Home

Symantec White Paper - Implementing Solaris Zones with Veritas

1. From CVM Master node vxdctl c mode gt vxdg s init mydata_dg mydata_dg01 hitachi_usp vm0_083e gt vxassist g mydata_dg make mydata_vol 1g mydata_dg0O1 gt mkfs F vxfs dev vx rdsk mydata_dg mydata_vol From Node1 gt mkdir zones endzone mydata From Node2 gt mkdir zones calzone mydata From CVM Master Important Note For this scenario the commands for adding a Cluster Mount to your VCS configuration will require that you first choose the same entry for the MountPoint argument for the cfsmntadm command followed by modifying the Mount Point Attribute using the hares modify flag so that it is localized to each Host gt cfsmntadm add mydata_dg mydata_vol zones mydata zone_zfs_SG node1 suid rw node2 suid rw This will add a CVMVoldg and CFSMount Resource to the local_vxfs_zone_SG service group gt hares local cfsmount1 MountPoint The naming convention of cfsmount is the default naming scheme for adding CFSMount resources You may choose to modify the resource name offline by editing the main cf or via the copy paste function on the VCS Java GUI gt hares modify cfsmount1 MountPoint zones calzone mydata sys node gt hares modify cfsmount1 MountPoint zones endzone mydata sys node2 gt hares modify cfsmount1 NodeList node node2 gt hares modify cfsmount1 Primary node gt haconf dump makero gt hagrp online local_vxfs_zone_SG any 47 Implementing Solaris Zones
2. bash 3 2 zonecfg z zonename calzone No such zone configured se create to begin configuring a new zone zonecfg zonename gt create b zonecfg zonename gt set zonepath zones zonename base zonecfg zonename gt set autoboot false zonecfg zonename gt commit zonecfg zonename gt verify zonecfg zonename gt exit Note that these steps must be performed on each node in the cluster where the local zone is configured to run You can optionally export the Zone configuration to a file and configure any subsequent nodes by copying the configuration file to remaining nodes and using the following commands Node1 gt zonecfg z zonenmame export f zonename cfg gt scp zone cfg node2 tmp Node2 gt zonecfg z zonename f tmp zonename cfg 11 Implementing Solaris Zones with Veritas Cluster Server from Symantec and Veritas Cluster file System High Availability 6 0 from Symantec SUMMARY Controlling applications running in Solaris local zones using Veritas Cluster Server is a relatively straight forward process There are however a number of configuration options that can be chosen to achieve a variety of different results Whether your goal is to reduce physical server footprint or simply increase utilization of the available host side resources Solaris zones are very effective means to achieve these milestones Combining however Cluster File System High Availability and Solaris local zones provide
3. Zs Resource View yrts_vea_cfs_int_cfsmount2 on em gi S Ss P Fe E vrts_vea_cfs_int_cfsmo Nodes A g Figure 13 Application Storage and Zone Root Resources Step 2 Configure amp Install each Zone and Incorporate CFS LOFS File Systems C Define the Zone on each corresponding Physical Node Example From Node1 gt zonecfg z twilightzone create b set zonepath zones twilightzone base set autoboot false add fs set dir mydata set special zones twilightzone mydata set type lofs end commit verify exit gt zoneadm z twilightzone install gt zoneadm z twilightzone boot gt zlogin C twilightzone 31 Implementing Solaris Zones with Veritas Cluster Server from Symantec and Veritas Cluster file System High Availability 6 0 from Symantec Follow Prompts for configuring zone Once complete you can use the keystroke to exit the console and return to the OS prompt gt cp etc resolv conf zones dangerzone base root etc gt cp etc nsswtich zones dangerzone base root etc This procedure will only work consistently for shared IP zones D Export Zone Configuration to Node2 From Nodel gt zonecfg z twilightzone export f tmp twilightzone cfg gt scp tmp twilightzone node2 tmp From Node2 gt zonecfg z twlightzone f tmp twilightzone cfg Step 3 Configure Service Groups to support Zones and for Application Failover For Verita
4. D Configure Individual Zone IP Addresses Node1 amp Node2 gt ifconfig interface plumb gt ifconfig interface 1 Address netmask netmask zone zonename up Example node1 gt ifconfig bgeO 1 plumb gt ifconfig bgeO 1 10 10 10 1 netmask 255 255 255 0 zone dangerzone up Example node2 gt ifconfig bgeO 1 plumb gt ifconfig bgeO 1 10 10 10 2 netmask 255 255 255 0 zone redzone up 26 Implementing Solaris Zones with Veritas Cluster Server from Symantec and Veritas Cluster file System High Availability 6 0 from Symantec E Run the hazonesetup script on each cluster node Nodel gt hazonesetup t g local_vxfs_zone_SG r localzone_ZONE z dangerzone u z_vcs_dangerzone p password s node Node2 gt hazonesetup t g local_vxfs_zone_SG r localzone_ZONE z redzone u z_vcs_redzone p password I s node2 F Add Zone IP and NIC Resource to Service Group Add NIC Resource From Either Node gt hares add zone_NIC NIC local_vxfs_zone_SG gt hares modify Zone NIC Device bgeO gt hares modify zone_NIC Critical 0 gt hares modify zone_NIC Enabled 1 Add IP Resource From Either Node gt hares add zone_IP IP local_vxfs_zone_SG gt hares modify IP Critical 0 gt hares modify IP Device bgeO gt hares local IP Address gt hares modify IP Address 10 10 10 1 sys node gt hares modify IP Address 10 10 10 2 sys node2 gt hares modify IP NetMask
5. Implementing Solaris Zones with Veritas Cluster Server from Symantec and Veritas Cluster file System High Availability 6 0 from Symantec Who should read this paper The audience for this document is system administrators who need best practices for configuring local zones in Veritas Cluster Server and for systems maintenance when local zones are placed under Veritas Cluster Server control Moreover provided within this guide will be the steps necessary to establish a Cluster File System High Availability based cluster topology for both zone root and application data file systems Confidence in a connected world symantec Implementing Solaris Zones with Veritas Cluster Server from Symantec and Veritas Cluster file System High Availability 6 0 from Symantec Content ACKNOWLEDGEMENTS wisi is 5 6 5 6 6 9 Y spa enya sate TRER Jaa a ward RR aie N ER K RER dais RN E R ARR RRR E R KERN RRR HREH TR 1 DOCUMENT CONSIDERATIONS cscs 67 RETS see ad S K Boia Rinses R RRC Geko avn ERRER R ale X RRR cla ace ale RER ins ayes eas 1 INTRODUCTION lt cssccisicsschscin cs sigvsasatara etia aca i tetera doe EE Sewn pie 678 a at ETE EE evga EE E wo p Si auarape E E biwla oma 1 INTERACTION BETWEEN VERITAS CLUSTER SERVER AND SOLARIS LOCAL ZONES x K K cece cece cece cece cece scenes 3 VERITAS CLUSTER SERVER AGENT FRAMEWORK CHANGES FOR 5 1 xx s x x lt x x e x eee 3 VERITAS CLUSTER SERVER RESOURCE amp RESOURCE TYPE C
6. Symantec the Symantec Logo and the N Checkmark Logo are trademarks or registered and contact numbers please 350 Ellis St lt br gt Mountain View trademarks of Symantec Corporation or its affiliates in the U S and other countries Other names may be visit our website CA 94043 USA lt br gt 1 650 527 trademarks of their respective owners 2 2013 21281112 8000 lt br gt 1 800 721 3934 lt br gt www symantec com
7. gt chmod 700 zones gold_image_zone gt zoneadm z gold_image_zone install gt zoneadm z gold_image_zone detach gt umount zones gold_image_zone Creating the Zone Root Snapshot Preparing Zone Root Volume for Snapshot Operations One Time Operation gt vxsnap g cfs_zroot_dg prepare cfs_zroot_vol Establishing the Cache Volume One Time Operation gt vxasssit g cfs_zroot_dg make cfs_zroot_cachevol alloc cfs_zroot_dg01 Creating the Cache Object One Time Operation gt vxmake g cfs_zroot_dg cache cfs_zroot_cachobj cachevolname cfs_zroot_cachevol autogrow on Establishing Point in Time Space Optimized Snapshot gt vxsnap g cfs_zroot_dg make source cfs_zroot_vol newvol newzonel_zroot_vol cache cfs_zroot_cacheobj Verifying Snapshot 55 Implementing Solaris Zones with Veritas Cluster Server from Symantec and Veritas Cluster file System High Availability 6 0 from Symantec gt vxsnap g cfs_zroot_dg list gt vxcache g cfs_zroot_dg stat Creating a New Zone Using a Space Optimized Snapshot The following commands make the assumption that the snapshot will be mounted as traditional VxFS file system as opposed to Cluster File System That said Cluster File System can certainly be used as long as the DG is imported with Shared flag Much of the commands displayed here can be modified and automated based on your particular environment What is important to note is that the time to clone a whole
8. 255 255 240 0 gt hares modify IP Enabled 1 Configure Resource Dependencies From Either Node gt hares link localzone_ZONE cfsmount1 gt hares link localzone_ZONE zoneroot_MNT gt hares link zone_IP localzone_ZONE gt hares link zone_IP Zone NIC gt haconf dump makero 27 Implementing Solaris Zones with Veritas Cluster Server from Symantec and Veritas Cluster file System High Availability 6 0 from Symantec f demo_vcs6 TG xi FA Resource View local_vxfs_zone_SG on all systems Sw Figure 11 Completed Resource Dependency View Once the Zone service group is completed you will want establish a separate failover service group to support your application s The dependency between the Application Parent service group and the Zone Child service group should be Online Local Firm Please follow the VCS solutions guides for specifics on how to configure the supported applications for failover gt hagrp link my_application_SG local_vxfs_zone_SG online local firm https sort symantec com documents doc_details sfha 6 0 Solaris ProductGuides 28 Implementing Solaris Zones with Veritas Cluster Server from Symantec and Veritas Cluster file System High Availability 6 0 from Symantec Option 3 Cluster File System Zone Root Shared and Cluster File System for Application Data LOFS In this scenario a single zone will be configured with its root file system on shared storag
9. Server Important note ZFS by default provides an automatic import service based on the following Solaris 10 SMF entry svc network shares group zfs What this service provides is the ability for Solaris 10 to upon creation of a Zpool or reboot import the storage pool and subsequently automount the ZFS file systems without using the etc vfstab file As a best practice when using Veritas Cluster Server to manage ZPools or VxVM volumes for that matter as part of Service group you will want to configure the mount point attribute for each pool and file system to legacy so as to allow Veritas Cluster Server to manage the import export and mount operations exclusively To identify what the mountpoint attribute for a particular ZPool or ZFS file system is by running the following command gt zfs get mountpoint poolname or poolname file system name The necessary syntax to configure this attribute is provided in the following sections 35 Implementing Solaris Zones with Veritas Cluster Server from Symantec and Veritas Cluster file System High Availability 6 0 from Symantec Example 1 Local Zone Root with Shared Data File Systems Direct Mount ZFS In this example ZFS will be used to configure both the Zone root and application data file systems Although not required it is still a best practice to place local zone not just shared zone root file systems on SAN attached storage for increased flexibility With the relea
10. VCS zones cfs zroot base Restoring Entire Zone Root File System This process will upon completion delete the storage checkpoint You may choose to create a new checkpoint before remounting the Zone root file system From Either Node gt opt VRTSvxfs sbin fsckpt_restore dev vx dsk cfs_zroot_dg cfs_zroot_vol cfs_zroot_ckpt dev vx dsk cfs_zroot_dg cfs_zroot_vol cfs_zroot_ckpt ctime November 5 2012 09 45 33 PM PST mtime November 5 2012 09 45 33 PM PST flags largefiles removable UX vxfs fsckpt_restore WARNING V 3 24640 Any file system changes or storage checkpoints made after November 5 2012 09 45 33 PM PST will be lost Restore the filesystem from storage checkpoint cfs_zroot_ckpt ynq y Yes UX vxfs fsckpt_restore INFO V 3 23760 File system restored from cfs_zroot_ckpt Unfreezing Service Group From Either Node gt hagrp unfreeze Service Group Mounting Zone Root File System and Booting Zone From Either Node gt hares online Storage Service Group any gt hares online Zone Application Service Group sys node Zone Provisioning with Cluster File System and Space Optimized Snapshots Cluster File System Users may choose to provision zones simply by creating a golden unconfigured image on a CVM volume and subsequently use the SOSS facility to establish new read write volumes Once created these newly minted zone images may be configured as they would with any traditional zonecf
11. anid E EEE aan E eis grand aca asa yo 0 Wieser 12 Option 1 VxFS Zone Root Local and Cluster File System for Application Data Direct Mount e e e eee 12 Option 2 VxFS Zone Root Local and Cluster File System for Application Data LOFS cc ccc cece cece ceceeeeceees 21 Option 3 Cluster File System Zone Root Shared and Cluster File System for Application Data LOFS c eee eeeee 29 Appendix A Veritas Cluster Server Local Zones and Native Volume Management ccc c x x x x x x e e e eee 35 Appendix B Zone Provisioning with Cluster File System and FlashSnap ccc x x x x cece cece cence eee ee ee esses seseeees 51 Appendix C Applying Patches to Systems with Zones Under Veritas Cluster Server Control ccc cece eee 60 Implementing Solaris Zones with Veritas Cluster Server from Symantec and Veritas Cluster file System High Availability 6 0 from Symantec ACKNOWLEDGEMENTS would like to take this opportunity to acknowledge the contributing members of the SAMG team specifically Eric Hennessey and James Jax Jackson Both of whom without would have not been able to complete the refresh of this document This content provided and the best practices put forth here were not exclusively developed in the vacuum of a lab environment would like to extend my sincere appreciation to not only the customer contributors but the Symantec Product Management En
12. base root etc resolv conf gt cp etc nsswtich zones dangerzone base root etc This procedure will only work consistently for shared IP zones From Node2 gt zonecfg z redzone create b set zonepath zones redzone base set autoboot false add fs set dir m data set special zones redzone mydata set type lofs end commit verify exit gt zoneadm z redzone install gt zoneadm z redzone boot gt zlogin C redzone Follow Prompts for configuring the zone Once complete you can use the keystroke to exit the console and return to the OS prompt gt cp etc resolv conf zones redzone base root etc gt cp etc nsswtich zones redzone base root etc This procedure will only work consistently for shared IP zones Step 3 Configure Service Groups to support Zones and for Application Failover For Veritas Cluster Server to effectively monitor applications that reside inside of a local zone Veritas Cluster Server must be configured to allow communications to occur from the global zone to the local zone and vice versa This is established using a combination of the halogin and hazonesetup commands Please note that any resources configured with the Critical flag set to O are done so for initial deployment purposes only HAZONESETUP This utility establishes the following configuration settings e Creates a new or modifies an existing Service Group to support the ContainerInfo Attributes e Defines wh
13. configured on VxFS CFS a snapshot either full sized or space optimized can be taken in order to provide a duplicate image of the zone root that may be tested upon off host Combined with the checkpoint process previously outlined the effect of patching a zone can be mitigated and the process successfully vetted without disrupting any production applications The following procedures in conjunction with the aforementioned Flashsnap operations will provide the high level steps necessary for testing Solaris patches for both local and shared zones either via the space optimized or full sized snapshot method One important consideration is that each zone has been configured without an IP address assigned This allows the snapshot zone to be booted concurrently with the original environment so as to avoid causing an IP address conflict Space Optimized Method using Cluster File System The main caveat to this process is the existence of a node within the cluster that is eligible to be patched i e one that does not currently host production applications This is most commonly available in a N to N cluster topology where a specific node can be used as the patch target Preferably choose a node not in the systems list for the application or zone service groups 1 Establish cache volume and cache object for zone root volume 2 Create Space Optimized Snapshot of zone root volume w Export the configuration for the original zone to designated pa
14. modify mydata_ZFS FsckOpt n modify mydata_ZFS Enabled 1 Add IP Resource From Either Node gt hares gt hares gt hares gt hares gt hares gt hares add application_IP IP zfs_application_SG modify IP Critical 0 modify IP Device bgeO modify IP Address 10 10 10 6 modify IP NetMask 255 255 240 0 modify IP Enabled 1 Adding Application Resource Please consult the corresponding solutions guide for Veritas Cluster Server on how to configure monitoring for a specific application type Implementing Solaris Zones with Veritas Cluster Server from Symantec and Veritas Cluster file System High Availability 6 0 from Symantec https sort symantec com documents doc_details sfha 6 0 Solaris ProductGuides Configure Resource Dependencies From Either Node gt hares link mydata_ZFS mydata_Zpool gt hares link my_application_APP application_IP gt hares link my_appliaction_APP mydata_ZFS gt haconf dump makero gt hagrp online zfs_application_SG sys node T demo_vcs6 Simulator gj x H Resource View 2fs_application_SG on all systems P A Once the Zone service group is configured you will want configure separate resources for your application s The dependency between the zone application parent service group and the storage child service group should be Online Local Firm Please follow the Veritas Cluster Server solutions guides for specifics on how to conf
15. root zone and assign various configuration parameters is substantially reduced gt mkdir zones newzonel gt mount F vxfs dev vx dsk cfs_zroot_dg newzonel_zroot_vol zones newzonel gt chmod 700 zones newzonel gt zonecfg z newzonel newzonel No such zone configured Use create to begin configuring a new zne zonecfg newzonel1 gt create b zonecfg newzone1 gt set zonepath zones newzone1 zonecfg newzone1 gt set autoboot false zonecfg newzonel gt commit zonecfg newzone1 gt verify zonecfg newzonel gt exit gt zoneadm z newzonel attach F gt zoneadm z newzonel boot gt zlogin C newzonel a Follow the normal process for configuring Zone OS Zone Provisioning with Cluster File System and Full Size Instant Snapshots For those environments that require an even greater degree of zone flexibility or the use of off host processing the full sized snapshot option will satisfy both of these requirements By incorporating Snapshot Mirrors to an either active zone root volume or an unconfigured golden image users will be able to provision new zones or stage a cloned zone root for purpose of patch certification The following procedure outlines the steps required to establish a full sized instant snapshot for the purposes of zone provisioning Preparing a Zone for Snapshot Provisioning The process to provision new zones using full sized instant snapshots will require duplicate storage capacity Zone root vol
16. shared IP zones Bring the IP Resources Online From Either Node gt hares online myzone_IP sys Node gt hares online myzone_IP sys Node2 From Either Cluster Node Important note When configuring Zones to with Local Zone root file systems and shared Data mounts you must create a parallel service group and localize the Pool Mount Point and ContainerInfo attributes A Run the hazonesetup script on each cluster node From Node1 gt hazonesetup t g zone_zfs_SG r myzone_ZONE z calzone u z_vcs_calrzone p password I s node From Node2 gt hazonesetup t g zone_zfs_SG r myzone_ZONE z endzone u z_vcs_endzone p password I s node2 Bringing Service Group Online gt hares modify myzone_ZONE Enabled 1 gt hares link cfsmount1 myzone_ZONE gt hagrp online zone_zfs_SG any 49 Implementing Solaris Zones with Veritas Cluster Server from Symantec and Veritas Cluster file System High Availability 6 0 from Symantec f demo_vcs6 E Resource View zone_zfs_SG on all systems Figure 19 ZFS and Cluster File System Coexistence Once the Zone service group is configured you will want configure separate resources for your application s The dependency between the zone application parent service group and the storage child service group should be Online Local Firm Please follow the Veritas Cluster Server solutions guides for specifics on how to configure the supported a
17. zone_NIC Critical 0 gt hares modify zone_NIC Enabled 1 Add IP Resource From Either Node gt hares add zone_IP IP local_vxfs_zone_SG gt hares modify IP Critical 0 gt hares modify IP Device bgeO gt hares local IP Address gt hares modify IP Address 10 10 10 1 sys node gt hares modify IP Address 10 10 10 2 sys node2 gt hares modify IP NetMask 255 255 240 0 gt hares modify IP Enabled 1 19 Implementing Solaris Zones with Veritas Cluster Server from Symantec and Veritas Cluster file System High Availability 6 0 from Symantec Configure Resource Dependencies From Either Node gt hares link cfsmount1 localzone_ZONE gt hares link localzone_ZONE zoneroot_MNT gt hares link zone_IP localzone_ZONE gt hares link zone_IP Zone NIC gt haconf dump makero T demo_vcs6 x U F Resource View local_vxfs_zone_SG on all systems BPH Ao arange P B z local vas 20n Nodet Figure 8 Completed Resource Dependency View Once the Zone service group is completed you will want establish a separate failover service group to support your application s The dependency between the Application Parent service group and the Zone Child service group should be Online Local Firm Please follow the Veritas Cluster Server solutions guides for specifics on how to configure the support applications for failover gt hagrp link my_application_SG local_vxfs
18. Cluster Server user used for password less communication a Populate AutoStartList for the group Configure a parallel service group If you do not specify the option a failover service group is created by default 5 Systems A comma separated list of systems where the zone service group need to be configured for example sy51 5y52 sys53 18 Implementing Solaris Zones with Veritas Cluster Server from Symantec and Veritas Cluster file System High Availability 6 0 from Symantec A Configure Individual Zone IP Addresses Node1 amp Node2 gt ifconfig interface plumb gt ifconfig interface 1 Address netmask netmask zone zonename up Example node1 gt ifconfig bge0 1 plumb gt ifconfig bge0 1 10 10 10 1 netmask 255 255 255 0 zone dangerzone up Example node2 gt ifconfig bge0 1 plumb gt ifconfig bge0 1 10 10 10 2 netmask 255 255 255 0 zone redzone up B Run the hazonesetup script on each cluster node Nodel gt hazonesetup t g local_vxfs_zone_SG r localzone_ZONE z dangerzone u Z Veritas Cluster Server_dangerzone p password s node Node2 gt hazonesetup t g local_vxfs_zone_SG r localzone_ZONE z redzone u Z Veritas Cluster Server_redzone p password s node2 C Add Zone IP and NIC Resource to Service Group Add NIC Resource From Either Node gt hares add zone_NIC NIC local_vxfs_zone_SG gt hares modify Zone NIC Device bgeO gt hares modify
19. EM HIGH AVAILABILITY Important note When configuring a local zone which will be under Veritas Cluster Server control there is only one required deviation from the default build the zone must be configured so that it will not boot when the system is started as Veritas Cluster Server will be responsible for controlling the boot execution for the zone Installing Non Global Zones The available options for when initially configuring a Zone are too numerous to cover for scope of this document However for the sake of context and understanding the necessary steps the basic zonecfg and zoneadm commands required are detailed below Additionally it is assumed that all LUNS have been presented to the corresponding nodes as well as formatted and initialized for use with Volume Manager Please consult the Storage Foundation Administrators guide for details on preparing storage for use with VxVM https sort symantec com documents doc_details sfha 6 0 Solaris ProductGuides Implementing Solaris Zones with Veritas Cluster Server from Symantec and Veritas Cluster file System High Availability 6 0 from Symantec Example Zone Configuration Overview Throughout this document the example of a two node Veritas Cluster Server cluster will be used The basic configuration is as follows Cluster Nodes node amp node2 Local zones with VxFS Cluster File System root dangerzone and redzone Shared Zone with Cluster File System Root twilightzo
20. HANGES lt x x x x lt x x K K K eee 4 Pacl ALEN Touren draie Pace snusearie tena data le scape tadus a yeube die Setters cele nceubseetast autenennlie g E alesse A dai Mita Bier a at Gcoue aden Regen bun ware dea ea Sere 4 Other bundled agents sr occas ond meee Hee ek EE EEA EEI wee eee AN E Ada ee es daa ae ets 5 CLUSTER FILE SYSTEM HIGH AVAILABILITY OVERVIEW cece cece K eee 5 FILE SYSTEM SUPPORT FOR ZONES 7 4075 RER obra sa vn laa ad sve destaias poe TRR ERRER ERRER drag acon R TRT el nissan 6 Sample Cluster Topology 2 3 60 cessaci ar eee KK R Rd Rd R ee de POM EEA SEW Sed RR RR RR RRR awa dda ee R RAR RRR RR K NA 7 Best Practices For Local Zone Configuration in Veritas Cluster Server cence ene t ete eee e nee nee 7 CONFIGURING A LOCAL ZONE TO WORK WITH CLUSTER FILE SYSTEM HIGH AVAILABILITY 0 eee cece cece cece eee cerns 8 installing Non Global Zones lt 0 2 4 N 6 R 09 aed wa dice a ed Sa RAS SEEM A ALS REREAD NG MEAG alee oS aed oe SEO NS eed 8 Example Zone Configuration Overview ene n en ene ene n ene ene n ete et ee tenet eenees 9 Defining The Non Global Zone xac S c ad e boa ec bee eee Ree dR La OA REDHEAD Ce EEEE eed eb bee desea eeu e ede eee ES Ewa awe 9 Sample Zone Configuration gR arg R 64066 06 te eee be ee cae ee ee aed ee De Sade Chota ender ERE bed E Wud Ee ebesaeeabeed 11 SUMMARY conesa 015 606 octgc sic 0 s rinse sw nin o d 6 oie o e 0 fo bia ols Dia Si aw SUR rw hea ag abo ona
21. _vxfs_zone_SG gt hares modify Zone NIC Device bgeO 33 Implementing Solaris Zones with Veritas Cluster Server from Symantec and Veritas Cluster file System High Availability 6 0 from Symantec gt hares modify zone_NIC Critical 0 gt hares modify zone_NIC Enabled 1 Add IP Resource From Either Node gt hares add zone_IP IP local_vxfs_zone_SG gt hares modify IP Critical 0 gt hares modify IP Device bgeO gt hares local IP Address gt hares modify IP Address 10 10 10 1 sys node gt hares modify IP Address 10 10 10 2 sys node2 gt hares modify IP NetMask 255 255 240 0 gt hares modify IP Enabled 1 Configure Resource Dependencies From Either Node gt hares link localzone_ZONE cfsmount1 gt hares link localzone_ZONE zoneroot_MNT gt hares link zone_IP localzone_ZONE gt hares link zone_IP zone NIC gt haconf dump makero rr SOt ORACLE Gy Resource View failover_zone_app_SG on all systems These Resources must be added separately Figure 14 Example Failover Application Service Group Application Resources Shown 34 Implementing Solaris Zones with Veritas Cluster Server from Symantec and Veritas Cluster file System High Availability 6 0 from Symantec Once the Zone service group is completed you will want establish a separate failover service group to support your application s The dependency between the Application Parent servic
22. _zone_SG online local firm https sort symantec com documents doc_details sfha 6 0 Solaris ProductGuides 20 Implementing Solaris Zones with Veritas Cluster Server from Symantec and Veritas Cluster file System High Availability 6 0 from Symantec Option 2 VxFS Zone Root Local and Cluster File System for Application Data LOFS In this configuration each zone root will reside on local storage while the application data will reside on Cluster File System The unit of failover will be the application and as such two separate service groups will be required One parallel SG for the zone and Cluster File System file system and one SG for the applications and virtual P s i kae ww IP Application Failover ae Online Local Firm Ip Zone Zone E B Mout S CFSMount NIC App Data aE Cero Seri Volume aaas gt App Data Zone Root Parallel DiskGroup gt Service Group Zone Root Figure 9 Service Group Dependency for LOFS Mount Cluster File System Step 1 Configure Zone Root and Application Disk Groups Volumes and File Systems A Create Disk Groups on each host for the Zone Root Node1 amp Node2 From Node gt vxdg init dangerzone_zroot_dg dangerzone_zroot_dg01 hitachi_usp vm0_083f From Node2 gt vxdg init redzone_zroot_dg redzone_zroot_dg01 hitachi_usp vm0_0840 B Create Volumes and File Systems f
23. a robust flexible and scalable solution that addresses a variety of use cases These include reducing the downtime associated with a number of dally operational tasks providing increased levels of availability as well as allowing for enterprise wide visibility to all Solaris IBM VMware Red Hat and Microsoft included virtual environments through the use of VERITAS Operations Manager Furthermore Cluster File System High Availability will dramatically reduce the recovery times associated with planned as well as unplanned outages No true cluster solution would be complete unless it addresses both Local High Availability as well as DR considerations Veritas Cluster Server does precisely that by allowing for DR personalization for use with replication of the zone root IT Enterprises can now confidently replicate their Solaris virtual environments without concern for the often tedious and manual process of preparing a Zone for use in a new Data center For more information about this feature please refer to the Veritas Cluster Server Bundled Agents Reference Guide on Symantec Operation Readiness Tools Cluster File System High Availability is comprehensive solution which improves upon that native Solaris zone framework This is done so by providing the same High Availability Disaster Recovery feature set that users have been implementing for years with Veritas Cluster Server Having adapted them in to the virtual context this inevitably makes the
24. adoption of containers a much easier process Option 1 VxFS Zone Root Local and Cluster File System for Application Data Direct Mount In this configuration each zone root will reside on local storage while the application data will reside on Cluster File System The unit of failover will be the application and as such two separate service groups will be required One parallel SG for the zone and Cluster File System file system and one failover SG for the applications and virtual IP s 12 Implementing Solaris Zones with Veritas Cluster Server from Symantec and Veritas Cluster file System High Availability 6 0 from Symantec i Application op Application Failover gt Service Group Online Local Firm ie CFSMount 4 IP L Zone App Data a ze J CvMvoldg gt Mount 7 Volume Sone root 2 pa X Era Parallel DiskGroup Service Group Zone Root _ 2 Figure 6 Service Group Dependency for Direct Mount Cluster File System Step 1 Configure Zone Root and Application Disk Groups Volumes and File Systems It is the assumption of this document that all LUN s have been formatted labeled and initialized for use with Volume Manager prior to executing the following steps A Create Disk Groups on each host for the Zone Root Node1 amp Node2 From Node gt vxdg init dangerzone_zroot_dg dangerzone_zro
25. ares gt hares gt hares gt hares modify zoneroot_ modify zoneroot_ modify zoneroot_ DG Critical 0 DG StartVolumes 0 DG StopVolumes 1 ocal zoneroot_DG DiskGroup modify zoneroot_ modify zoneroot_ modify zoneroot_ Add Volume Resource DG DiskGroup dangerzone_zroot_dg sys node DG DiskGroup redzone_zroot_dg sys node2 DG Enabled 1 gt hares add zoneroot_DG DiskGroup local_vxfs_zone_SG gt hares gt hares gt hares gt hares gt hares gt hares gt hares gt hares modify zoneroot_VOL Critical 0 ocal zoneroot_VOL Volume modify zoneroot_VOL Volume dangerzone_zroot_vol sys node modify zoneroot_VOL Volume endzone_zroot_vol sys node2 ocal zoneroot_VOL DiskGroup modify zoneroot_VOL DiskGroup dangerzone_zroot_dg sys node modify zoneroot_VOL DiskGroup endzone_zroot_dg sys node2 modify zoneroot_VOL Enabled 1 gt hares link zoneroot_VOL zoneroot_DG Add Mount Resource gt hares add zoneroot_MNT Mount local_vxfs_zone_SG gt hares gt hares gt hares gt hares gt hares gt hares gt hares gt hares gt hares gt hares modify zoneroot_ modify zoneroot_ modify zoneroot_ modify zoneroot_ modify zoneroot_ modify zoneroot_ modify zoneroot_ modify zoneroot_ NT Critical 0 ocal zoneroot_MNT MountPoint NT MountPoint zones dangerzone base sys node NT MountPo
26. at the Zone or Container name be defined as part of the Zone agent itself users are no longer required to define that attribute Rather the Zone agent now simply references the string assigned to the Name key within the ContainterInfo service group attribute amp Important note With the release of Solaris 10 8 07 Update 4 Sun introduced two new functions to the online and offline operation associated with local zones Users can now choose to either attach or detach a zone as part of the process for moving or decoupling a zone between physical global zones This plays a significant role when determining the patching strategy particularly if zones are the predominant virtual environment being deployed When a zone is shutdown it is the default behavior for packages that contain the following attribute to boot all offline Zones prior to adding the package SUNW_ALL_ZONES true This setting is defined within the package s pkginfo file However should the zone be detached the pkgadd operation will NOT consider the zone eligible for booting The following output of the zoneadm command identifies whether a zone is attached detached as well as online or offline gt zoneadm list civ ID NAME STATUS PATH BRAND IP 0 global running native shared 1 calzone running zones calzone base native shared When a zone is fully booted zoneadm will report its status as running while a zone that is attached but not booted will repo
27. ces for system patching Prior to Solaris 10 8 07 the lack of an attach detach function made this a rather challenging process That said with the incorporation of the detach option and more recently the Update on Attach feature Solaris zone patching is a decidedly easier operation than was the case with earlier iteration of Solaris 10 As mentioned previously in this guide the decision of where to place the zone root file system will have a direct effect on the process you must follow to apply system patches For those environments where the zone itself remains resident while the application is the unit of failover consistency between the global zone and local zones is less of a challenge However what will be a hurdle to overcome is when an individual zone requires that you apply a particular patch but due to nature of the zone i e being defined only on single host you will be unable to move the environment to an alternate host so as to avoid testing in production Moreover a roll back target would not exist to restore the zone root in the event of corruption resulting for a failed package or patch install Due to the way in which the patchadd and pkgadd commands work from the global zone the net result of this operation will be that all attached local zones are patched concurrently This is not the most ideal circumstance for an untested patch to be installed on a production system Alternatively for those zones where the root file system is
28. e The application data will also reside on Cluster File System which can be presented as LOFS or direct mount The unit of failover will be the entire zone and as such the Zone and Application resources will be configured in a same service group The Zone Root and application storage resources will be configured as a parallel service group Application IP Zone and App Failover Service Group J Online Local Firm CFSMount CFSMount Zone Root App Data CVMVoldg CVMVoldg Zone Root App Data Parallel Service Group Figure 12 Service Group Dependency for LOFS CFS Step 1 Configure Zone Root and Application Disk Groups Volumes and File Systems A Create Shared Disk Group Volume and File System for the Zone Root From Master Node gt vxdg s init twilightzone_zroot_dg dangerzone_zroot_dg01 hitachi_usp vm0_083f 29 Implementing Solaris Zones with Veritas Cluster Server from Symantec and Veritas Cluster file System High Availability 6 0 from Symantec Example Master Node gt vxassist g twilightzone_zroot_dg make twilightzone _zroot_vol 5g twilightzone _zroot_dg01 gt mkfs F vxfs dev vx rdsk twilightzone _zroot_dg twilightzone _zroot_vol gt mkdir p zones twilightzone base gt mount f vxfs o cluster suid rw dev vx dsk twilightzone _zroot_dg twilightzone _zroot_vol zones twilightzone base gt chmod 700 zones twilightzone base Node 2 gt mkdi
29. e following considerations Implementing Solaris Zones with Veritas Cluster Server from Symantec and Veritas Cluster file System High Availability 6 0 from Symantec e Required Unit of Failover Zone or Application e Tolerance for Patching Outages e User Account Control s Is there a need to reduce the number of Individual OS environments s SAN Impact for increase in virtual environments Sample Cluster Topology Application Failover Application CFSHA 6 0 Solaris 10 Global Zone Gt Solaris 10 Global Zone o Wome yam Figure 4 Cluster Topology when using CFSHA with Solaris Zones Best Practices For Local Zone Configuration in Veritas Cluster Server Choosing the Location for the Zone Root File System Veritas Cluster Server supports placement of the zone root file system on either shared or local storage The advantage of placing the zone root file system on shared storage is that the zone installation must be performed only once This does have a disadvantage however when it comes time to apply system patches This topic is described in more detail in Appendix C There are number of additional considerations when deciding the placement of zone root file systems For the purposes of this document however the most important scenarios to consider will be the following 1 Shared Zone Root with Zone and Application Failover 2 Local Zone Root with Application Failover Only In the first example a non globa
30. e group and the Zone Child service group should be Online Local Firm Please follow the VCS solutions guides for specifics on how to configure the supported applications for failover gt hagrp link my_application_SG local_vxfs_zone_SG online local firm https sort symantec com documents doc_details sfha 6 0 Solaris ProductGuides Appendix A Veritas Cluster Server Local Zones and Native Volume Management In this section the following topics will be covered e Overview of the Veritas Cluster System support for ZFS e Configuring ZFS pools and File Systems e Incorporating ZFS resources into failover and Parallel service groups e Sample configurations e Configuring ZFS and Cluster File System coexistence Please note that although minimal CLI syntax will be provided to show the best practice for integrating ZFS into Veritas Cluster Server this section will not cover in any great detail ZFS operations or administration Please consult Oracle documentation for and additional ZFS related inquiries Veritas Cluster Server Support for ZFS Bundled with Veritas Cluster Server are two agents that support native ZFS components the ZPool and Mount agents The ZPool agent was introduced to provide control for the import and export operations associated with ZFS storage pools To support ZFS file systems the Mount agent has been modified to include ZFS as an available file system type Configuring ZFS components for use with Veritas Cluster
31. e one application cannot effect the operation of another One feature that Solaris zones do share with the Virtual Machine concept is the notion of representing not only multiple instances but versions of the same operating system This however is achieved through the use of branded Zones as opposed to individual kernel isolation Implementing Solaris Zones with Veritas Cluster Server from Symantec and Veritas Cluster file System High Availability 6 0 from Symantec Disk Storage and File Systems Figure 1 Relationship of local zones to the global zone For more information on zones and resource management refer to the Oracle reference guide System Administration Guide Oracle Solaris Containers Resource Management and Oracle Solaris Zones September 2010 no 817 1592 http docs oracle com cd E19253 01 817 1592 817 1592 pdf Beginning with Veritas Cluster Server version 4 1 Veritas added support for Solaris zones in a clustered environment This allows a system administrator to start stop and monitor an application within the confines of a local zone and failover zones between systems in a cluster The intent of this document is to provide systems administrators the information needed to correctly configure local zones in a Veritas Cluster Server cluster and to provide best practices for systems maintenance when local zones are placed under Veritas Cluster Server control Moreover provided within this guide will be
32. e the overall demands placed upon the resources of an individual host by the multiple instances of our traditional monitoring agents When coupled with the use of Cluster File System you are able to dramatically reduce not only the overhead on the system but the time required to failover an application or zone between physical hosts With version 6 0 support for the Zone agent was added to the list of IMF aware resource types Please note whereas with 5 1SP1 enabling IMF was a manual procedure starting in 6 0 all IMF supported resource types will have this setting enabled by default With the enabling of the IMF attribute for the Zone agent a nearly 80 percent performance improvement with regard to CPU load was achieved for online zones To see a list of the agents supported by IMF you can run the following command gt haimfconfig display Implementing Solaris Zones with Veritas Cluster Server from Symantec and Veritas Cluster file System High Availability 6 0 from Symantec VERITAS CLUSTER SERVER RESOURCE amp RESOURCE TYPE CHANGES Along with agent framework modifications several core agents have been modified or created that supports the ContainerOpts attribute settings These are described below Zone Agent With the release of versions 5 1 of Veritas Cluster Server the bundled Zone agent has been modified to support the ContainerInfo Service Group attribute Whereas previous releases of Veritas Cluster Server required th
33. ed to track updates to both the original volume and its snapshot Reads of blocks that have not been modified are satisfied from original volume Auxiliary application Production application Hie Writes to ETTE Reads of blocks Bild a urr snapshot that have been write blocks go jy 11 modified are original into cache satisfied from volume volume cache object Cache volume Original volume backing store for cache LIT TIT TTT Cache ob Kal Ra p 11 changed regions Snapshot DCO ea a Volume DC iL Laez Cranes La Tracks updated First update to a region Tracks updated regions regions of Causes Copy on write to of original volume snapshot cache object Storage Checkpoints Storage Checkpoints are allocate on first write snapshots of a VxFS file system They are entirely contained within the file system block space and require no external volumes or other structures A Storage Checkpoint consists of a map that initially contains indirect pointers to the locations of data and metadata in the original file system The map becomes populated with pointers to prior images of data and metadata as it is updated in the original file system over time 51 Implementing Solaris Zones with Veritas Cluster Server from Symantec and Veritas Cluster file System High Availability 6 0 from Symantec 3 i i N w Modified block snapshot area Se L Volume containing Auxiliary application 3 Other amp applicatio
34. endencies From Either Node gt hares link myzone_ZONE zoneroot_MNT gt hares link zoneroot_MNT myzone_ZPool gt hares link myzone_IP myzone_ZONE gt hares link myzone_IP zone NIC gt haconf dump makero From Both Cluster Nodes Define the Zone on each corresponding Physical Node Example From Node1 gt zonecfg z calzone create b set zonepath zones calzone base set autoboot false 39 Implementing Solaris Zones with Veritas Cluster Server from Symantec and Veritas Cluster file System High Availability 6 0 from Symantec end commit verify exit gt zoneadm z calzone install gt zoneadm z calzone boot gt zlogin C calzone Follow Prompts for configuring zone Once complete you can use the keystroke to exit the console and return to the OS prompt gt cp etc resolv conf zones calzonezone base root etc gt cp etc nsswtich zones calzone base root etc This procedure will only work consistently for shared IP zones Configure the Second Zone on Node 2 From Node 2 gt zonecfg z endzone create b set zonepath zones endzone base set autoboot false end commit verify exit gt zoneadm z endzone install gt zoneadm z endzone boot gt zlogin C endzone Follow Prompts for configuring zone Once complete you can use the keystroke to exit the console and return to the OS prompt gt cp etc resolv conf zones endzone base root e
35. eritas Cluster file System High Availability 6 0 from Symantec From Either Node gt hagrp add zfs_application_SG gt hares add mydata_ZPool Zpool zfs_application_SG gt hares gt hares gt hares gt hares gt hares gt hares gt hares modify mydata_ZPool Critical 0 modify mydata_ZPool ChkZFSMounts 0 modify myzone_ZPool PoolName appdata_zpool sys node1 node2 modify myzone_ZPool ZoneResName myzone_ZONE modify myzone_ZPool Enabled 1 modify zoneroot_MNT FSType zfs modify zoneroot_MNT FsckOpt n gt hares online myzone_ZPool sys node Important note you may choose to have the ZPool agent check the status of the ZFS mounts with the ChkZFSMounts attribute This implies the Mount agent is not part of the Service Group This document assumes the Mount agent will be configured Adding th e Application Mount and Network Resources Adding Mount Resource From Either Node gt hares add mydata_ZFS Mount myzone_SG gt hares gt hares gt hares gt hares gt hares gt hares gt hares gt hares gt hares modify mydata_ZFS Critical 0 modify mydata_ZFS CreateMntPt 1 ocal mydata_ZFS MountPoint modify mydata_ZFS MountPoint zones calzone base root mydata sys node modify mydata_ZFS MountPoint zones endzone base root mydata sys node2 modify mydata_ZFS BlockDevice appdata_zpool appdata modify mydata_ZFS FSType zfs
36. erly recognizes the newly created ZPools you must first export the zone root and data storage pool as shown above therefore allowing Veritas Cluster Server to successfully bring the ZPool resource online Adding the Zone Root ZPool Resources Adding ZPool Resource From Either Node gt hagrp add local_vxfs_zone_SG gt hagrp modify myzone_SG SystemList node1 0 node2 1 gt hagrp modify myzone_SG Parallel 1 gt hares add myzone_ZPool Zpool myzone_SG gt hares modify myzone_ZPool Critical 0 gt hares modify myzone_ZPool ChkZFSMounts 0 gt hares local myzone_ZPool PoolName gt hares modify myzone_ZPool PoolName calzone_zroot sys node gt hares modify myzone_ZPool PoolName endzone_zroot sys node2 gt hares modify myzone_ZPool ZoneResName myzone_ZONE gt hares modify myzone_ZPool Enabled 1 gt hares online myzone_ZPool sys node gt hares online myzone_ZPool sys node2 Important note you may choose to have the ZPool agent check the status of the ZFS mounts with the ChkZFSMounts attribute This implies the Mount agent is not part of the Service Group This document assumes the Mount agent will be configured Adding the Zone Root Mount and Network Resources Adding Mount Resource From Either Node gt hares add zoneroot_MNT Mount myzone_SG gt hares modify zoneroot_MNT Critical 0 gt hares modify zoneroot_MNT CreateMntPt 1 gt hares local zoneroot_MNT Mo
37. ether the SG is of the parallel or failover variety e Creates a new resource of Type Zone in the aforementioned SG 25 Implementing Solaris Zones with Veritas Cluster Server from Symantec and Veritas Cluster file System High Availability 6 0 from Symantec e Optionally creates new Individual users for Veritas Cluster Server authentication from the non global zone Otherwise a predefined user can be chosen e Established Authentication between the non global zone and Global Zone halogin From the Global Zone gt hazonesetup t sg name r res_name z zone_name u user name p password a I s sys1 sys2 Where the Values are t Updates the password for the YCS zone user g sg_name Name of the zone service group to be created in YCS configuration r res_name Name of the zone resource to be created in VCS configuration z zane_name Name of the zone that is configured on the system u user_name Name of the VCS user used for password less communication between the local zone and the global zone If no username is specified the default username is used p password Password for the YCS user used for password less communication Populate AutoStartList for the group Configure a parallel service group If you do not specify the option a failover service group is created by default 5 Systems A comma separated list of systems where the zone service group need to be configured for example sys1 sys2 sys3
38. evalent than in the case of zone provisioning Whether you want to expedite the deployment of zones implement an instantaneous roll back procedure or reduce the storage consumed by an increased number of temporary OS instances i e for dev test and lab operations Storage Foundation has the corresponding snapshot type to support these efforts Use Case 1 Data Corruption and Patch Roll Back In this scenario a Solaris zone whose root resides on VxFS will have a Storage Check Point associated to its root file system 52 Implementing Solaris Zones with Veritas Cluster Server from Symantec and Veritas Cluster file System High Availability 6 0 from Symantec Use Case 2 New Zone Provisioning with CFS and Space Optimized Snapshots SOSS In this scenario an unconfigured Solaris zone root file system residing on VxFS will have one or more Space Optimized Snap Shots SOSS attached to the parent volume Use Case 3 New Zone provisioning off host with full split mirror snapshots In this scenario a Solaris zone whose root file system is on VxFS will have a full sized snapshot mirror attached to its root volume Use Case 4 Pre Staging Patches for Local Zone Details for this option are covered in Appendix C Zone Root Roll Back using VxFS Checkpoints The purpose of this configuration is to provide instantaneous rollback for any changes made to a zone root file system These may include failed application installs corrupted packages
39. figuration z zone_name Name of the zone that is configured on the system u user_name Name of the VCS user used for password less communication between the local zone and the global zone If no username is specified the default username is used p password Password for the VCS user used for password less communication a Populate AutoStartList for the group Configure a parallel service group If you do not specify the option a failover service group is created by default 5 systems A comma separated list of systems where the zone service group need to be configured for example sys1 sys2 sys3 E Configure Individual Zone IP Addresses Node1 amp Node2 gt ifconfig interface plumb gt ifconfig interface 1 Address netmask netmask zone zonename up Example node1 gt ifconfig bgeO 1 plumb gt ifconfig bgeO 1 10 10 10 1 netmask 255 255 255 0 zone dangerzone up Example node2 gt ifconfig bgeO 1 plumb gt ifconfig bgeO 1 10 10 10 2 netmask 255 255 255 0 zone redzone up F Run the hazonesetup script on each cluster node Nodel gt hazonesetup t g local_vxfs_zone_SG r localzone_ZONE z dangerzone u z_vcs_dangerzone p password I s node Node2 gt hazonesetup t g local_vxfs_zone_SG r localzone_ZONE z redzone u z_vcs_redzone p password I s node2 G Add Zone IP and NIC Resource to Service Group Add NIC Resource From Either Node gt hares add zone_NIC NIC local
40. g and zoneadm command syntax As stated previously the only infrastructure prerequisite for this operation is the existence of available space in the disk group so as to create the SOSS cache volume Preparing a Zone for Snapshot Provisioning 54 Implementing Solaris Zones with Veritas Cluster Server from Symantec and Veritas Cluster file System High Availability 6 0 from Symantec The process to provision new zones using Cluster File System and Space Optimized Snapshots is relatively straightforward Zone root volumes are treated no differently than any other data source when configuring these types of snapshots The process begins with installing but not configuring a new zone root image This will act as the golden image source for all other zone roots and therefore only requires to be installed one time Create Source Zone Root File System Please refer to the section that previously covered creating disk groups volumes and file systems for use with zone roots Preparing a Golden Image Zone One time Operation From Either Node gt zonecfg z gold_image_zone gold_image_zone No such zone configured Use create to begin configuring a new zone zonecfg gold_image_zone gt create b zonecfg gold_image_zone gt set zonepath zones gold_image_zone zonecfg gold_image_zone gt set autoboot false zonecfg gold_image_zone gt commit zonecfg gold_image_zone gt verify zonecfg gold_image_zone gt exit
41. ginal file system Export the configuration for the original zone to the designated patch testing node Configure the snapshot zone for the new global using the exported cfg file from above zonecfg z testzone f testzone cfg Remove any LOFS data entries from the zone configuration IO 0 I ae Qe IS a Attach and boot snapshot zone 2 Apply any new packages or patches to the snapshot zone Re LA Halt snapshot zone on alternate host and detach it N Deport the new DG from the alternate host ES Import and join snapshot dg and source dg on production host TS Reattach refresh zone root snapshot mirror 61 Implementing Solaris Zones with Veritas Cluster Server from Symantec and Veritas Cluster file System High Availability 6 0 from Symantec About Symantec Symantec protects the world s information and is a global leader in security backup and availability solutions Our innovative products and services protect people and information in any environment from the smallest mobile device to the enterprise data center to cloud based systems Our world renowned expertise in protecting data identities and interactions gives our customers confidence in a connected world More information is available at www symantec com or by connecting with Symantec at go symantec com socialmedia For specific country offices Symantec World Headquarters Copyright 0 2075 symantec Corporation Allright reserved
42. gineering and Support teams all of whose willingness to indulge my requests made this effort possible You know who you are DOCUMENT CONSIDERATIONS The comments and best practices included within this document assume a certain set of prerequisites to support the published capabilities and functionality Along with those environmental factors this guide contains additional content specific considerations These include but are not limited to the following e The operating environment is Solaris 10 Update 8 or higher Scalable Processor Architecture SPARC s All Cluster Nodes are physical hosts and not Oracle Virtual Machine and Logical Domain VM s LDoms That said most of the same considerations will also apply to configuring Zones relative to LDoms as opposed to physical servers All Zone examples in this document will be of the Whole Root variety For details on Whole Root vs Sparse Zones please consult Oracle Zone Administration Guide http docs oracle com cd E19455 01 817 1592 fgotf index html s Although both are supported with Veritas Cluster Server from Symantec it is the assumption of this document that the Shared IP as opposed to the Exclusive IP model will be configured for network connectivity in each example e Zetabyte File System ZFS specific considerations will be addresses exclusively in the Appendix Veritas Cluster Server Local Zones and Native Volume Management s Although not explici
43. igure the supported applications for failover https sort symantec com documents doc_details sfha 6 0 Solaris ProductGuides Example 2 Local Zone Root with Shared Data File Systems CFS Direct Mount In this example the Zone root will reside on ZFS however the application data will use Cluster File System mounts As with the previous example the zone root file systems will be place on SAN attached storage for flexibility Also as is the case with scenario 1 the Zones themselves will not be shared but rather the application will be the unit of failover 43 Implementing Solaris Zones with Veritas Cluster Server from Symantec and Veritas Cluster file System High Availability 6 0 from Symantec 7 Rr 7 N cfsMount IP ae E Sar Mount zi Figure 18 Service Group Dependency for Direct Mount Cluster File System From Cluster Node 1 gt zpool create calzone_zroot c1t50060E8005631F51d1 creates ZPool on specified device gt zfs create calzone_zroot calzone_root creates ZFS File System for Zone Root gt zfs set mountpoint legacy calzone_zroot disables automount auto import gt zfs set mountpoint legacy calzone_zroot calzone_root disables automount gt zfs list NAME USED AVAIL REFER MOUNTPOINT calzone_zroot 152K 9 60G 32K legacy calzone_zroot calzone_root 31K 9 60G 31K legacy gt zpool export calzone_zroot From Cluster Node 2 I
44. int zones redzone base sys node2 ocal zoneroot_MNT BlockDevice NT BlockDevice dev vx dsk dangerzone_zroot_dg dangerzone_zroot_vol sys node NT BlockDevice dev vx dsk redzone_zroot_dg redzone_zroot_vol sys node2 NT FSType vxfs NT FsckOpt n NT Enabled 1 gt hares link zoneroot_MNT zoneroot_VOL 14 Implementing Solaris Zones with Veritas Cluster Server from Symantec and Veritas Cluster file System High Availability 6 0 from Symantec D Create CVM Disk Groups Volumes and File Systems for Application Data From either Cluster File System node gt vxdg s init mydata_dg mydata_dgO1 hitachi_usp vm0_083e gt vxassist g mydata_dg make mydata_vol 1g mydata_dg01 gt mkfs F vxfs dev vx rdsk mydata_dg mydata_vol From Node1 gt mkdir zones dangerzone mydata gt mount f vxfs o cluster suid rw dev vx dsk mydata_dg mydata_vol zones dangerzone mydata From Node2 gt mkdir zones redzone mydata gt mount f vxfs o cluster suid rw dev vx dsk mydata_dg mydata_vol zones redzone mydata 15 Implementing Solaris Zones with Veritas Cluster Server from Symantec and Veritas Cluster file System High Availability 6 0 from Symantec From Either Node Important Note For this scenario the commands for adding a cluster mount to your Veritas Cluster Server configuration will require that you first choose the same entry for the MountPoint argument for the cfsmntadm command fol
45. l Zpool zone_zfs_SG gt hares modify myzone_ZPool Critical O gt hares modify myzone_ZPool ChkZFSMounts 0 gt hares local myzone_ZPool PoolName gt hares modify myzone_ZPool PoolName calzone_zroot sys node gt hares modify myzone_ZPool PoolName endzone_zroot sys node2 gt hares modify myzone_ZPool ZoneResName myzone_ZONE gt hares modify myzone_ZPool Enabled 1 45 Implementing Solaris Zones with Veritas Cluster Server from Symantec and Veritas Cluster file System High Availability 6 0 from Symantec gt hares online myzone_ZPool sys node gt hares online myzone_ZPool sys node2 Important note you may choose to have the ZPool agent check the status of the ZFS mounts with the ChkZFSMounts attribute This implies the Mount agent is not part of the Service Group This document assumes the Mount agent will be configured Adding the Zone Root Mount and Network Resources Adding Mount Resource From Either Node gt hares add zoneroot_MNT Mount zone_zfs_SG gt hares modify zoneroot_MNT Critical 0 gt hares modify zoneroot_MNT CreateMntPt 1 gt hares local zoneroot_MNT MountPoint gt hares modify zoneroot_MNT MountPoint zones endzone base sys node2 gt hares modify zoneroot_MNT MountPoint zones calzone base sys node gt hares local zoneroot_MNT BlockDevice gt hares modify zoneroot_MNT BlockDevice endzone_zroot endzone_root sys node2 gt hare
46. l zone will be essentially portable and subject to patching considerations when detached and attached to a different global zone physical host This can be advantageous if your concern is in the number of operating user environments and keeping them to a minimum Also with the use of Cluster File System you can completely remove the delay associated with deporting and importing the zone root and application storage One additional advantage here is in the ability to create a Golden Zone Image one that has not been Implementing Solaris Zones with Veritas Cluster Server from Symantec and Veritas Cluster file System High Availability 6 0 from Symantec configured and use Flashsnap to provision new environments nearly instantaneously This is covered in Appendix B The disadvantage however is that you must be cognizant of the patch level on each global zone and avoid patch drift between the global and non global zones To some degree this has been addressed with the addition of the Update on Attach function from Solaris 10 It is important though to note that this feature is not supported by the Veritas Cluster Server Zone agent and will not covered in this document The second scenario will mitigate all patching concerns as the zone root itself is not portable but rather localized to the physical host This can still be a SAN attached VxFS file system but not one that would be deported on imported on a regular basis The Veritas Cluster Ser
47. lowed by modifying the Mount Point attribute using the hares modify flag so that it is localized to each Host gt cfsmntadm add mydata_dg mydata_vol zones mydata local_vxfs_zone_SG node1 suid rw node2 suid rw This will add a CVMVoldg and CFSMount resource to the local_vxfs_zone_SG service group gt hares local cfsmount1 MountPoint The naming convention of cfsmount is the default naming scheme for adding Cluster File System Mount resources You may choose to modify the resource name offline by editing the main cf or via the copy paste function on the Veritas Cluster Server Java GUI gt hares modify cfsmount1 MountPoint zones dangerzone mydata sys node gt hares modify cfsmount1 MountPoint zones redzone mydata sys node2 gt hares modify cfsmount1 NodeList node node2 gt hares modify cfsmount1 Primary node gt haconf dump makero gt hagrp online local_vxfs_zone_SG any E demo_vcs6 Pile x F Resource View local_vxfs_zone_SG on all systems PUR R auto Arrange D ve J local_vxfs_zon twet32n sol30 uxrt32n sol28 Figure 7 Resource Dependency View 16 Implementing Solaris Zones with Veritas Cluster Server from Symantec and Veritas Cluster file System High Availability 6 0 from Symantec Step 2 Configure amp Install each Zone A Define each individual Zone on its corresponding Physical Node From Node1 gt zonecfg z dangerzone create b set zo
48. lows for the configuration of application resources that can failover between hosts irrespective of whether the environment is virtualized or not This by definition enables P 2 V Physical to Virtual cluster topologies that otherwise would be in many cases unnecessarily complex to deploy To effectively support this modification the Veritas Cluster Server 5 1 agent framework has introduced a new service group attribute entitled ContainerInfo This attribute when configured defines the Type Zone or Wpar Name and Enabled 0 1 or 2 Designed to work in conjunction with the newly added resource type attribute ContainerOpts these two configurable attributes provide a comprehensive and flexible framework to support all manner of application and zone monitoring requirements For specific details on the various attribute settings please refer to the Storage FoundationHigh Availability Virtualization Guide for Solaris http www symantec com business support resources sites BUSINESS content live DOCUMENTATION 5000 D0C5405 en_US sfha_virtualization_60_sol pdf It is also worth noting that with the release of Veritas Cluster Server 5 1 SP1 Symantec has introduced the concept of an Intelligent Monitoring Framework IMF or Kernel Level Monitoring The purpose of this evolution past the traditional polling agent methodology is intended to first provide for instantaneous notification of application state changes and subsequently reduc
49. mplementing Solaris Zones with Veritas Cluster Server from Symantec and Veritas Cluster file System High Availability 6 0 from Symantec gt zpool create endzone_zroot c2t50060E8005631F41d1 lt Only visible to Node 2 gt zfs create endzone_zroot endzone_root gt zfs set mountpoint legacy endzone_zroot gt zfs set mountpoint legacy endzone_zroot endzone_root gt zfs list NAME USED AVAIL REFER MOUNTPOINT endzone_zroot 4 47G 5 13G 32K legacy endzone_zroot endzone_root 4 47G 5 13G 4 47G legacy gt zpool export endzone_zroot Important note To ensure that VCS properly recognizes the newly created ZPools you must first export the zone root and data storage pool as shown above therefore allowing VCS to successfully bring the ZPool resource online From Either Cluster Node Important note When configuring Zones to with Local Zone root file systems and shared Data mounts you must create a parallel service group and localize the Pool Mount Point and ContainerInfo attributes Build The Zone Service Group gt haconf makerw gt hagrp add zone_zfs_SG gt hagrp modify zone_zfs_SG SystemList nodel 0 node2 1 gt hagrp modify zone_zfs_SG Parallel 1 Important note You will want to leave the Zone resource disabled prior to actually configuring and installing the Zone itself Adding the Zone Root ZPool Resources Adding ZPool Resource From Either Node gt hares add myzone_ZPoo
50. nd vice versa This is established using a combination of the halogin and hazonesetup commands Please note that any resources configured with the Critical flag set to 0 are done so for initial deployment purposes only HAZONESETUP This utility establishes the following configuration settings e Creates a new or modifies an existing Service Group to support the ContainerInfo Attributes e Defines whether the SG is of the parallel or failover variety e Creates a new resource of Type Zone in the aforementioned SG e Optionally creates new Individual users for Veritas Cluster Server authentication from the non global zone Otherwise a predefined user can be chosen e Established Authentication between the non global zone and Global Zone halogin From the Global Zone gt hazonesetup t sg name r res_name z zone_name u user name p password a I s sys1 sys2 Where the Values are t Updates the password for the Veritas Cluster Server zone user gsg_name Name of the zone service group to be created in Veritas Cluster Server configuration rres_name Name of the zone resource to be created in Veritas Cluster Server configuration zzone_name Name of the zone that is configured on the system u user_name Name of the Veritas Cluster Server user used for password less communication between the local zone and the global zone If no username is specified the default username is used p password Password for the Veritas
51. ne Local zones with ZFS root calzone and endzone VxFS Zone root file system zones zonename base ZFS Zone root file system zones zonename base Cluster File System LOFS Application volumes mounted at zones zonename data_mnt The Examples given in this document will cover the following 5 scenarios s VxFS Root for Local Zone with Cluster File System Direct Mount for Application Data App Failover VxFS Root for Local Zone with Cluster File System LOFS for Application Data App Failover s Cluster File System Root for Shared Zone with Cluster File System LOFS for Application Data Zone App Failover Appendix A e ZFS Root for Local Zone with ZFS Direct Mount for Application Data Zone and App Failover e ZFS Root for Local Zone with Cluster File System Direct Mount for Application Data Application Failover In order to support the scenarios above for application and Zone failover the following considerations must be noted e The application service group failover must be dependent upon the Zone root and data storage service group using online local firm e When configuring Zones for application failover you must localize the Containerlnfo Service Group attributes Name Type and Enabled e The Mountpoint attribute for the Cluster File SystemMount data resource must also be localized Node 1 zones ZoneA root mydata Node 2 zones ZoneB root mydata s Once the zone is brought online with Direct Mount file sys
52. ne_zroot_vol sys node2 gt hares modify zoneroot_MNT FSType vxfs gt hares modify zoneroot_MNT FsckOpt n gt hares modify zoneroot_MNT Enabled 1 gt hares link zoneroot_MNT zoneroot_VOL Bring Service Group Online gt hagrp online local_vxfs_zone_SG any D Create CVM Disk Groups Volumes and File Systems for Application Data From CVM Master node vxdctl c mode gt vxdg s init mydata_dg mydata_dg01 hitachi_usp vm0_083e gt vxassist g mydata_dg make mydata_vol 1g mydata_dg0O1 gt mkfs F vxfs dev vx rdsk mydata_dg mydata_vol From Node1 gt mkdir zones dangerzone mydata From Node2 gt mkdir zones redzone mydata From CVM Master Important Note For this scenario the commands for adding a cluster mount to your Veritas Cluster Server configuration will require that you first choose the same entry for the MountPoint argument for the cfsmntadm command followed by modifying the Mount Point Attribute using the hares modify flag so that it is localized to each host gt cfsmntadm add mydata_dg mydata_vol zones mydata local_vxfs_zone_SG node1 suid rw node2 suid rw This will add a CVMVoldg and CFSMount Resource to the local_vxfs_zone_SG service group gt hares local cfsmount1 MountPoint The naming convention of cfsmount is the default naming scheme for adding CFSMount resources You may choose to modify the resource name offline by editing the main cf or via the copy
53. nepath zones dangerzone base set autoboot false commit verify exit gt zoneadm z dangerzone install gt zoneadm z dangerzone boot gt zlogin C dangerzone Follow Prompts for configuring zone Once complete you can use the keystroke to exit the console and return to the OS prompt gt cp etc resolv conf zones dangerzone base root etc gt cp etc nsswtich zones dangerzone base root etc This procedure will only work consistently for shared IP zones From Node2 gt zonecfg z redzone create b set zonepath zones redzone base set autoboot false commit verify exit gt zoneadm z redzone install gt zoneadm z redzone boot gt zlogin C redzone Follow Prompts for configuring the zone Once complete you can use the keystroke to exit the console and return to the OS prompt gt cp etc resolv conf zones redzone base root etc gt cp etc nsswtich zones redzone base root etc This procedure will only work consistently for shared IP zones 17 Implementing Solaris Zones with Veritas Cluster Server from Symantec and Veritas Cluster file System High Availability 6 0 from Symantec Step 3 Configure Service Groups to support Zones and for Application Failover For Veritas Cluster Server to effectively monitor applications that reside inside of a local zone Veritas Cluster Server must be configured to allow communications to occur from the global zone to the local zone a
54. ns Write May read and Update IT file system mnt Virtual image vito Storage Y anes mntsnap1 Checkpoints 1 File system block space read from file File system block space s Create cman m system area snapshot 1 in Virtual image mntsnap 2 p mmni ak 2 File system block space data and Mount Storage metadata Checkpoints as file syst HL 12 K Copy volume data updated by application and update snapshot map Full Size Instant Snap Shots Split Mirror Split Mirror Snapshots are fully synchronized mirrors called plexes in VxVM terminology of volume contents that have been split from their original volumes at opportune moments and started as independent volumes either on or off host for application use Split Mirror Snapshot creation begins with the introduction of an additional mirror to a VxVM volume c 2 ia Volume DCO opy origina Original volume Snapshot volume contents be Cherie maged to added mirror IEP HEN ae 3 UE Pause UO and split iy ma oe AL mirror from R TL Additional mirror WE sali TU Additional mirror 4 Nay l Create data it LOT change objects UT Gila 1 UN IER il a Snapshot DCO Add mirror to volume ens Storage Foundation and FlashSnap Use Cases for Zones The storage resources required for Whole root Solaris Zones respond very well to disk based copy solutions and as such the use of snapshot facilities is very advantageous for a variety of operations Nowhere is this more pr
55. on gt vxsnap b g cfs_zroot_dg addmir cfs_zroot_vol alloc cfs_zroot_dg02 You will need to wait for the synchronization of the mirrors to complete before proceeding Taking a Full Sized Snapshot gt vxsnap g cfs_zroot_dg make source cfs_zroot _vol newvol newzone_vol plex cfs_zroot _vol 02 Splitting the Disk Group Private DG gt vxdg split cfs_zroot_dg newzone_zroot_dg newzone_vol gt vxdg deport newzone_zroot_dg gt vxdg import newzone_zroot_dg on alternate host Splitting the Disk Group Shared CVM DG gt vxdg split cfs_zroot_dg newzone_zroot_dg newzone_vol gt vxrecover b m g cfs_zroot_dg gt cfsdgadm move cfs_zroot_dg newzone_zroot_dg Mount New Snapshot Zone Root Alternate Host 57 Implementing Solaris Zones with Veritas Cluster Server from Symantec and Veritas Cluster file System High Availability 6 0 from Symantec gt mkdir p zones newzone base gt mount F vxfs dev vx dsk newzone_root_dg newzone_vol zones newzone base gt chmod 700 zones newzone base Mount New Snapshot Zone Root CFS gt haconf makerw gt mkdir p zones newzone base gt cfsmntadm add newzone_zroot_dg newzone_zroot_vol zones newzone base newzone_SG nodel rw suid gt hagrp online newzone_SG any gt haconf dump makero gt chmod 700 zones newzone base Creating a New Zone Using a Split Mirror Snapshot Once the Cluster File System mount is added to the cluster configuration
56. on options as shown above in Figure 2 Elaboration of the various VxFS Cluster File System and ZFS combinations are detailed in the following sections Figure 3 outlines the available combinations of the zone root path application data and the unit of failover App or Zone For those configurations where the unit of failover is the application the examples provided call for 2 independent zones neither of which will be decoupled or moved from their corresponding global zone ge LO C2 HOr EAD plication L a Local rare SAN or DAS ZFS Direct Mount rere ZFS or VxFS ZFS or VxFS ZFS or VxFS Cluster File System Cluster File System Shared SAN Local Storage SAN or DAS Shared SAN Local Storage SAN or DAS Shared SAN Shared SAN ZFS VxFS VxFS Cluster File System Custer File System Cluster File System Either Direct Mount Either Either Either Either zone Application zone Application zone Application Figure 3 Available combinations of Zone Root and Application Data Storage This option although allowing for the application to failover rather than the Zone calls for a single Zone Root to be concurrently booted on multiple Physical Nodes Global Zones and will not be covered in this guide Although the number of combinations for file system layout for both Zone root and data volumes is numerous the decision on which method to use in most cases is based on one or more of th
57. oot_DG DiskGroup dangerzone_zroot_dg sys node gt hares modify zoneroot_DG DiskGroup redzone_zroot_dg sys node2 gt hares modify zoneroot_DG Enabled 1 Add Volume Resource gt hares add zoneroot_DG DiskGroup local_vxfs_zone_SG gt hares modify zoneroot_VOL Critical 0 gt hares local zoneroot_VOL Volume gt hares modify zoneroot_VOL Volume dangerzone_zroot_vol sys node gt hares modify zoneroot_VOL Volume endzone_zroot_vol sys node2 gt hares local zoneroot_VOL DiskGroup gt hares modify zoneroot_VOL DiskGroup dangerzone_zroot_dg sys node gt hares modify zoneroot_VOL DiskGroup endzone_zroot_dg sys node2 gt hares modify zoneroot_VOL Enabled 1 gt hares link zoneroot_VOL zoneroot_DG Add Mount Resource gt hares add zoneroot_MNT Mount local_vxfs_zone_SG gt hares modify zoneroot_MNT Critical 0 gt hares local zoneroot_MNT MountPoint gt hares modify zoneroot_MNT MountPoint zones dangerzone base sys node gt hares modify zoneroot_MNT MountPoint zones redzone base sys node2 gt hares local zoneroot_MNT BlockDevice 22 Implementing Solaris Zones with Veritas Cluster Server from Symantec and Veritas Cluster file System High Availability 6 0 from Symantec gt hares modify zoneroot_MNT BlockDevice dev vx dsk dangerzone_zroot_dg dangerzone_zroot_vol sys node gt hares modify zoneroot_MNT BlockDevice dev vx dsk redzone_zroot_dg redzo
58. or each Zone Root Node1 amp Node2 Node 1 gt vxassist g dangerzone_zroot_dg make dangerzone_zroot_vol 5g dangerzone_zroot_dg01 gt mkfs F vxfs dev vx rdsk dangerzone_zroot_dg dangerzone_zroot_vol gt mkdir p zones dangerzone base gt mount f vxfs dev vx dsk dangerzone_zroot_dg dangerzone_zroot_vol zones dangerzone base gt chmod 700 zones dangerzone base Node 2 gt vxassist g redzone_zroot_dg make redzone_zroot_vol 5g redzone_zroot_dg01 gt mkfs F vxfs dev vx rdsk endzone_zroot_dg endzone_zroot_vol 21 Implementing Solaris Zones with Veritas Cluster Server from Symantec and Veritas Cluster file System High Availability 6 0 from Symantec gt mkdir p zones endzone base gt mount f vxfs dev vx dsk endzone_zroot_dg endzonezone_zroot_vol zones endzone base gt chmod 700 zones endzone base C Create Service Group and Resources for Zone Root From Either Node Create Service Group for Zone and Storage Resources gt haconf makerw gt hagrp add local_vxfs_zone_SG gt hagrp modify local_vxfs_zone_SG SystemList node 0 node2 1 gt hagrp modify local_vxfs_zone_SG Parallel 1 Add DiskGroup Resource gt hares add zoneroot_DG DiskGroup local_vxfs_zone_SG gt hares modify zoneroot_DG Critical 0 gt hares modify zoneroot_DG StartVolumes 0 gt hares modify zoneroot_DG StopVolumes 1 gt hares local zoneroot_DG DiskGroup gt hares modify zoner
59. or inadvertent deletion or modification of critical system files There is no infrastructure prerequisite as all data is stored with the context of the file system and therefore requires no additional storage space It is however important to note that during this procedure an outage will be necessary to perform any restore operations Creating the Roll Back Checkpoint Creating the Storage Checkpoint From Either Node gt fsckptadm create cfs_zroot_ckpt zones cfs zroot base Listing Checkpoints gt fsckptadm I list zones cfs zroot base zones cfs zroot cfs_zroot_ckpt ctime November 5 2012 09 45 33 PM PST mtime November 5 2012 10 07 52 PM PST flags largefiles removable mounted Mounting Unmounting Checkpoint To restore Individual Files From Either Node gt mkdir zones cfs zroot_ckpt gt mount F vxfs o cluster ckpt cfs_zroot_ckpt dev vx dsk cfs_zroot_dg cfs_zroot_vol cfs_zroot_ckpt zones cfs_zroot_ckpt gt umount zones cfs_zroot_ckpt Freezing All Corresponding Zone Service Groups Assuming Applications are already offline gt hagrp freeze Service Group 53 Implementing Solaris Zones with Veritas Cluster Server from Symantec and Veritas Cluster file System High Availability 6 0 from Symantec Unmounting Zone Root File System From Node1 gt opt VRTS bin umount 0 mntunlock VCS zones cfs zroot base From Node2 If using CFS gt opt VRTS bin umount 0 mntunlock
60. ot_dg01 hitachi_usp vm0_083f From Node2 gt vxdg init redzone_zroot_dg redzone_zroot_dg01 hitachi_usp vm0_0840 B Create Volumes and File Systems for each Zone Root Node1 amp Node2 Node 1 gt vxassist g dangerzone_zroot_dg make dangerzone_zroot_vol 5g dangerzone_zroot_dg01 gt mkfs F vxfs dev vx rdsk dangerzone_zroot_dg dangerzone_zroot_vol gt mkdir p zones dangerzone base gt mount f vxfs dev vx dsk dangerzone_zroot_dg dangerzone_zroot_vol zones dangerzone base gt chmod 700 zones dangerzone base Node 2 gt vxassist g redzone_zroot_dg make redzone_zroot_vol 5g redzone_zroot_dg0O1 gt mkfs F vxfs dev vx rdsk endzone_zroot_dg endzone_zroot_vol gt mkdir p zones endzone base gt mount f vxfs dev vx dsk endzone_zroot_dg endzonezone_zroot_vol zones endzone base gt chmod 700 zones endzone base C Create Service Group for Zone and Storage Resources 13 Implementing Solaris Zones with Veritas Cluster Server from Symantec and Veritas Cluster file System High Availability 6 0 from Symantec From Either Node Create Service Group for Zone and Storage Resources gt haconf makerw gt hagrp add local_vxfs_zone_SG gt hagrp modify local_vxfs_zone_SG SystemList node1 0 node2 1 gt hagrp modify local_vxfs_zone_SG Parallel 1 Add DiskGroup Resource gt hares add zoneroot_DG DiskGroup local_vxfs_zone_SG gt hares gt hares gt hares gt h
61. paste function on the Veritas Cluster Server Java GUI gt hares modify cfsmount1 MountPoint zones dangerzone mydata sys node1 gt hares modify cfsmount1 MountPoint zones redzone mydata sys node2 gt hares modify cfsmount1 NodeList node node2 gt hares modify cfsmount1 Primary node1 gt haconf dump makero gt hagrp online local_vxfs_zone_SG any 23 Implementing Solaris Zones with Veritas Cluster Server from Symantec and Veritas Cluster file System High Availability 6 0 from Symantec E demo_vcs6 Xx Pe Resource View local_vxfs_zone_SG on all systems Pink w as T P local_vxfs_zon uxrt32n sol30 uxrt32n sol28 Figure 10 Storage Resource Dependency View Step 2 Define and configure each individual Zone on its corresponding Physical Node From Node1 gt zonecfg z dangerzone create b set zonepath zones dangerzone base set autoboot false add fs set dir mydata set special zones dangerozne mydata set type lofs end commit verify exit gt zoneadm z dangerzone install gt zoneadm z dangerzone boot gt zlogin C dangerzone Follow Prompts for configuring zone Once complete you can use the keystroke to exit the console and return to the OS prompt 24 Implementing Solaris Zones with Veritas Cluster Server from Symantec and Veritas Cluster file System High Availability 6 0 from Symantec gt cp etc resolv conf zones dangerzone
62. pplications for failover https sort symantec com documents doc_details sfha 6 0 Solaris ProductGuides 50 Implementing Solaris Zones with Veritas Cluster Server from Symantec and Veritas Cluster file System High Availability 6 0 from Symantec Appendix B Zone Provisioning with Cluster File System and FlashSnap FlashSnap Overview A snapshot is a virtual image of the content of a set of data at the instant of creation Physically a snapshot may be a full complete bit for bit copy of the data set or it may contain only those elements of the data set that have been updated since snapshot creation The latter are sometimes referred to as copy on first write snapshots because space for data elements is added to the snapshot image only when the elements are updated overwritten for the first time in the original data set Storage Foundation copy on first write snapshots are called space optimized snapshots Space Optimized Instant Snap Shots SOSS A Space Optimized Instant Snapshot uses a disk based persistent cache object located on a cache volume to store prior images of original volume block regions as applications update blocks in them Cache objects map the addresses of updated block regions of the original volume to cache volume locations where their prior contents are saved In addition when VxVM creates a Space Optimized Instant Snapshot it creates DCOs for both the original volume and the snapshot The DCOs are us
63. r p zones twilightzone base gt mount f vxfs o cluster suid rw dev vx dsk twilightzone _zroot_dg twilightzone _zroot_vol zones twilightzone base gt chmod 700 zones twilightzone base gt haconf makerw gt cfsmntadm add twilightzone_zroot_dg twilightzone_zroot_vol zones twilightzone base shared_vxfs_zone_SG node1 suid rw node2 suid rw This will create the parallel service group local_vxfs_zone_SG with a CVMVoldg and CFSMount Resource B Create Shared Disk Group and CFS File Systems for Application Data Example From Master Node gt vxdg s init mydata_dg mydata_dg0O1 hitachi_usp vm0_0842 gt vxassist g mydata_dg make mydata_vol 1g mydata_dg0O1 gt mkfs F vxfs dev vx rdsk mydata_dg mydata_vol From Node1 gt mkdir zones twilightzone mydata gt mount f vxfs o cluster suid rw dev vx dsk mydata_dg mydata_vol zones twilightzone mydata From Node2 gt mkdir zones twilightzone mydata gt mount f vxfs o cluster suid rw dev vx dsk mydata_dg mydata_vol zones twilightzone mydata From Either Node gt cfsmntadm add mydata_dg mydata_vol zones twilightzone mydata shared_vxfs_zone_SG node1 suid rw node2 suid rw This will create the parallel service group local_vxfs_zone_SG with a CVMVoldg and CFSMount Resource 30 Implementing Solaris Zones with Veritas Cluster Server from Symantec and Veritas Cluster file System High Availability 6 0 from Symantec f demo_vcs6 Sisi x
64. r File System is the cornerstone of a highly available environment delivering faster recovery for business critical applications with reduced complexity and costs Compared to traditional single instance file system implementations Cluster File System significantly reduces application downtime and improves data access Because multiple servers share the same storage resources the length of time it takes to bring storage online is drastically reduced when an application fails Its high performance file system spans heterogeneous servers to provide concurrent access to data and enables faster failover of applications and databases such as Oracle SAP or Tibco Additionally Cluster File System provides the parallel data access necessary to allow for nearly instantaneous provisioning of new virtual environments These can include Solaris Zones Oracle VM s for SPARC LDoms as well as Red Hat KVM Virtual Machines Further elaboration of this topic will be covered later on in this document Implementing Solaris Zones with Veritas Cluster Server from Symantec and Veritas Cluster file System High Availability 6 0 from Symantec FILE SYSTEM SUPPORT FOR ZONES Zonecfg Direct mount LOFS Global Zone Figure 2 File System Options for Solaris Zones Depending on the requirement for the application being made highly available within the zone or simply just the zone itself Solaris 10 supports a wide variety of file system configurati
65. r error messages when attempting to configure DNS during the network configuration steps To mitigate this issue you can simply copy the etc resolv conf and nsswitch conf from the global zone to the local zone after the configuration is complete The final area of mention is the zonepath attribute The purpose for creating the additional base directory is to allow for a consistent mount point path for any subsequent LOFS or Cluster File System Direct Mount file systems that may be necessary This method is entirely optional and is done so purely for consistency purposes only 10 Implementing Solaris Zones with Veritas Cluster Server from Symantec and Veritas Cluster file System High Availability 6 0 from Symantec Sample Zone Configuration Creating the Local Root Zone Configuration for use with LOFS data Mounts bash 3 2 zonecfg z zonename calzone No such zone configured Use create to begin configuring a new zone zonecfg zonename gt create b zonecfg zonename gt set zonepath zones zonename base zonecfg zonename gt set autoboot false zonecfg zonename gt add fs zonecfg zonename fs gt set dir data zonecfg zonename fs gt set special zones zonename data zonecfg zonename fs gt set type lofs zonecfg zonename fs gt end zonecfg zonename gt commit zonecfg zonename gt verify zonecfg zonename gt exit Creating the Local Root Zone Configuration for use with Direct Mount File Systems
66. ration In order for Veritas Cluster Server to be effectively responsible for controlling the online and offline operations of the zone the auto boot attribute must be set to false The second area of note is the lack of any specific network information Veritas Cluster Server will make the appropriate IP address s available to the zone after it s booted using the IP Agent This is particularly useful when dealing with disaster recovery and the disaster recovery personalization features included with Veritas Cluster Server Keeping the resident configuration of the zone as generic as possible will allow for increased portability as well as simpler migrations There are however circumstances where as having the IP address available as part of the zonecfg process will be required One example would be the presence of NFS mounted file systems in the zone s etc vfstab You can however choose optionally to implement the Veritas Cluster Server Mount agent to manage the NFS file systems rather than the zone boot sequence itself It is also worth noting that should you choose to configure an IP address as part of the zonecfg process you will still want to use the IP Agent to monitor the status of the network address The is due to the default behavior that a Solaris Zones will remain booted even in the event that it s IP address is taken offline Important Note When you omit network information during the Zone build process you may encounte
67. rom Symantec Veritas Cluster Server brings IP resources online in the global zone by default If you want to bring these resources online inside the exclusive IP zone perform the following tasks e Make sure that the resource is in a service group that has valid Container nfo attribute value configured e Set the value of the ExclusivelPZone attribute to 1 Note The exclusive IP zone supports the IP and NIC networking agents For more information about these agents see the Veritas Cluster Server Bundled Agents Reference Guide http www symantec com business support resources sites BUSINESS content live DOCUMENTATION 5000 D0C5233 en_US vcs_bundled_agents_60_sol pdf Other bundled agents Along with the bundled Zone agent all remaining bundled agents have been modified to support the use of the ContainerOpts attribute Additionally the following replication agents have all been made container aware e Oracle Data Guard CLUSTER FILE SYSTEM HIGH AVAILABILITY OVERVIEW Cluster File System High Availability combines the industry leading Storage Foundation High Availability product set with the extended capability of Clustered Volume Manager and Cluster File System Together these tools allow for users to mount a Veritas File System VxFS on up to 64 Nodes concurrently for the supported UNIX and Linux operating systems Solaris AIX RedHat Enterprise Linux RHEL SUSE Linux and HP UX Veritas Storage Foundation Cluste
68. rt its status as being installed For those local zones that are shutdown and detached the zoneadm command with the civ flags will indicate which zones are in the configured state The entry points for the Zone agent online offline monitor amp clean are executed using a variation of the zoneadm boot zoneadm halt and zoneadm list commands Additionally the options for detach and attach f have been incorporated to the online and offline procedures Veritas Cluster Server by default will attach the zone on boot using the f force option and detach the zone when taken offline These setting however can be modified with the DetachZonePath and ForceAttach attributes If overridden this setting will cause the offline operation to leave the zone in the attached state Each of these settings is resource specific and therefore unique to each instance of the Zone Agent Networking Agents If you chose you can enable the attribute ExclusivelPZone for resources of type IP and NIC when these resources are configured to manage the IP and the NIC inside an exclusive IP zone This attribute is disabled by default The IP agent and the NIC agent assume the native zone behavior shared IP This however is no longer the case with Solaris 11 as exclusive IP vnic is the default setting Implementing Solaris Zones with Veritas Cluster Server from Symantec and Veritas Cluster file System High Availability 6 0 f
69. s modify zoneroot_MNT BlockDevice calzone_zroot calzone_root sys node gt hares modify zoneroot_MNT FSType zfs gt hares modify zoneroot_MNT FsckOpt n gt hares modify zoneroot_MNT Enabled 1 gt hares online zoneroot_MNT sys node gt hares online zoneroot_MNT sys node2 gt chmod 700 zones calzone base amp from node gt chmod 700 zones endzone base amp from node2 Add NIC Resource From Either Node gt hares add zone_NIC NIC zone_zfs_SG gt hares modify Zone NIC Device bgeO gt hares modify zone_NIC Critical 0 gt hares modify zone_NIC Enabled 1 Add IP Resource From Either Node gt hares add myzone_IP IP zone_zfs_SG gt hares modify IP Critical O gt hares modify IP Device bgeO gt hares local IP Address gt hares modify IP Address 10 10 10 4 sys node 46 Implementing Solaris Zones with Veritas Cluster Server from Symantec and Veritas Cluster file System High Availability 6 0 from Symantec gt hares modify IP Address 10 10 10 5 sys node2 gt hares modify IP NetMask 255 255 240 0 gt hares modify IP Enabled 1 Configure Resource Dependencies From Either Node gt hares link myzone_ZONE zoneroot_MNT gt hares link zoneroot_MNT myzone_ZPool gt hares link myzone_IP myzone_ZONE gt hares link myzone_IP zone NIC gt haconf dump makero Create CVM Disk Groups Volumes and File Systems for Application Data
70. s Cluster Server to effectively monitor applications that reside inside of a local zone Veritas Cluster Server must be configured to allow communications to occur from the global zone to the local zone and vice versa This is established using a combination of the halogin and hazonesetup commands Please note that any resources configured with the Critical flag set to O are done so for initial deployment purposes only HAZONESETUP This utility establishes the following configuration settings e Creates a new or modifies an existing Service Group to support the ContainerInfo Attributes e Defines whether the SG is of the parallel or failover variety e Creates a new resource of Type Zone in the aforementioned SG e Optionally creates new Individual users for Veritas Cluster Server authentication from the non global zone Otherwise a predefined user can be chosen Established Authentication between the non global zone and Global Zone halogin From the Global Zone gt hazonesetup t sg name r res_name z zone_name u user name p password a I s sys1 sys2 32 Implementing Solaris Zones with Veritas Cluster Server from Symantec and Veritas Cluster file System High Availability 6 0 from Symantec Where the Values are B Updates the password for the YCS zone user g sg_name Name of the zone service group to be created in VCS configuration rres_name Name of the zone resource to be created in VCS con
71. se of standalone DMP those users wishing to place zone root file systems on ZFS can do so by creating the underlying ZPool using Enclosure Based Naming EBN The following setting will need to be applied prior to configuring any of the ZFS components gt vxdmpadm settune dmp_native_support on App Service Group Figure 15 Service Group dependency for Local Zones using ZFS This would be the floating Virtual IP that is associated with the application and not the Zone itself This would be the zpool mount points for the Application Data You will want to set the path relative the Zone root For example if the Zone root is export Zones zonename and the application wants to access data then the mount point attribute would be export zones zonename root data You will also want to make sure the file system type is ZFS and that the actual Mountpoint property for the Zpool and FS itself is set to legacy 36 Implementing Solaris Zones with Veritas Cluster Server from Symantec and Veritas Cluster file System High Availability 6 0 from Symantec From Cluster Node 1 gt zpool create calzone_zroot c1t50060E8005631F51d1 creates ZPool on specified device gt zfs create calzone_zroot calzone_root creates ZFS File System for Zone Root gt zfs set mountpoint legacy calzone_zroot disables automount auto import gt zfs set mountpoint legacy calzone_zroot calzone_root disables automount gt zfs lis
72. ssic way of controlling cluster resources By introducing the Local Zone Container framework Veritas Cluster Server now has the ability to manage the start and stop of the local zones themselves as well as the constituent applications within the zones Effectively treating the zone as it would any other application resource After introducing the Zone agent to Veritas Cluster Server it became apparent that support for physical to virtual failovers would be a valuable configuration option Starting in version 5 1 a complete redesign of the Zone framework for Veritas Cluster Server was introduced to support this particular use case as well as simplify enablement VERITAS CLUSTER SERVER AGENT FRAMEWORK CHANGES FOR 5 1 The Veritas Cluster Server Agent Framework is a core set of functions that is compiled into every agent The agent is responsible for connecting with the Veritas Cluster Server engine HAD and carrying out core agent logic The Veritas Cluster Server agent framework first introduced the concept of Container Name with the release of 4 1 In subsequent releases specifically 5 1 Symantec redesigned how Cluster Server is configured to support virtual environments for both Solaris and Advanced Interactive Executive AIX With the release of version 5 1 the defining attributes for Container Name and Container Type i e Zone or WPar for AIX have been either modified or moved entirely to the Service Group layer By doing so this now al
73. t NAME USED AVAIL REFER MOUNTPOINT calzone_zroot 152K 9 60G 32K legacy calzone_zroot calzone_root 31K 9 60G 32K legacy gt zpool export calzone_zroot From Cluster Node 2 gt zpool create endzone_zroot c2t50060E8005631F41d1 Only visible to Node 2 gt zfs create endzone_zroot endzone_root gt zfs set mountpoint legacy endzone_zroot gt zfs set mountpoint legacy endzone_zroot endzone_root gt zfs list NAME USED AVAIL REFER MOUNTPOINT endzone_zroot 447G 5 13G 32K legacy endzone_zroot endzone_root 4 47G 5 13G 4 47G legacy gt zpool export endzone_zroot From Either Cluster Node Important note For this operation you will need to select a disk device that is visible to both cluster nodes as it will contain the shared application data gt zpool create appdata_zpool c2t50060E8005631F41d1 Shared LUN gt zfs create appdata_zpool appdata gt zfs set mountpoint legacy appdata_zpool gt zfs set mountpoint legacy appdata_zpool appdata gt zfs list NAME USED AVAIL REFER MOUNTPOINT appdata_zpool 152K 9 06G 32K legacy appdata_zpool appdata 31K 9 06G 31K legacy endzone_zroot 152K 9 06G 32K legacy endzone_zroot endzone_root 31K 9 06G 31K legacy 37 Implementing Solaris Zones with Veritas Cluster Server from Symantec and Veritas Cluster file System High Availability 6 0 from Symantec gt zpool export appdata_zpool Important note To ensure that Veritas Cluster Server prop
74. tc gt cp etc nsswtich zones endzone base root etc This procedure will only work consistently for shared IP zones Bring the IP Resources Online From Either Node gt hares online myzone_IP sys Node gt hares online myzone_IP sys Node2 40 Implementing Solaris Zones with Veritas Cluster Server from Symantec and Veritas Cluster file System High Availability 6 0 from Symantec From Either Cluster Node Important note When configuring Zones to with Local Zone root file systems and shared Data mounts you must create a parallel service group and localize the Pool Mount Point and ContainerInfo attributes A Run the hazonesetup script on each cluster node From Node gt hazonesetup t g local_zfs_zone_SG r localzone_ZONE z calzone u z_vcs_calrzone p password I s node From Node2 gt hazonesetup t g local_vxfs_zone_SG r localzone_ZONE z endzone u z_vcs_endzone p password s node2 Bringing Service Group Online gt hares modify myzone_ZONE Enabled 1 gt hagrp online myzone_SG f demo_vcs6 OE x EA Resource Yiew zone_zfs_SG on all systems H ats arrange P bs E E zone_zf Nodet Node2 Figure 16 Local Zone Service Group with ZFS Root Adding the Application Data Service Group and Resources Creating the Service Group and adding the ZPool Resource 41 Implementing Solaris Zones with Veritas Cluster Server from Symantec and V
75. tch testing node You will want to remove any LOFS definitions for data file systems on the snapshot zone Mount the Space Optimized Snapshot on the Patch Node using same path as with the source file system Define the temporary zone on the new host using zonecfg Boot the snapshot local zone Proceed with patch testing Shutdown and detach snapshot zone SO 00 h Oy OV oe Unmount snapshot root file system 60 Implementing Solaris Zones with Veritas Cluster Server from Symantec and Veritas Cluster file System High Availability 6 0 from Symantec At this point the snapshot can be refreshed at any time to create an up to date zone root for further testing Although these steps only cover the zone root volume application data can also be incorporated into this procedure very easily Choosing to do so will then allow for patch testing of application binaries in addition to the OS itself Full Sized Snapshot Method When choosing the full sized snapshot method users will be able to move the snapshot zone root off host to entirely independent node This can act as both a provisioning as well patch testing procedure Add second disk as a snapshot mirror to zone root volume Take snapshot of zone root Split disk group so as to create a new DG using only the snapshot zone root volume Deport and import the new DG to alternate node Mount the snapshot volume zone root on the target host using the same directory path as with the ori
76. tems this mount point will be visible only from within the Zone itself using df or by using the mount command from the global zone Defining The Non Global Zone While it s beyond the scope of this document to cover details on configuring a local zone some review of a zone s configuration is in order Local zone configurations are maintained in etc zones For each local zone configured on a host an entry exists in etc zones index and appears as follows calzone installed zones calzone base The components of the entry are zone name status and path to zone root separated by colons In the etc zones directory each local zone s configuration is stored in a file in XML format as lt zonename gt xml Figure 4 contains the entire zone configuration for our sample zone Implementing Solaris Zones with Veritas Cluster Server from Symantec and Veritas Cluster file System High Availability 6 0 from Symantec lt DOCTYPE zone PUBLIC Sun Microsystems Inc DTD Zones EN file usr share lib xml dtd zonecfg dtd 1 gt lt DO NOT EDIT THIS FILE Use zonecfg 1M instead gt lt zone name calzone zonepath zones dangerzone base autoboot false gt lt filesystem special zones dangerzone mydata directory mydata type lofs gt lt zone gt Figure 5 Local zone configuration file in XML format lt xml version 1 0 encoding UTF 8 gt There are several areas worth noting in the above configu
77. the steps necessary to establish a Cluster File System High Availability based cluster topology for both zone root and application data file systems Best practices noted in the document will be preceded by this symbol By the same token it will attempt to address configurations choices that should be carefully taken into consideration Such topics will be denoted with the following symbol What this document will not cover however is implementing Veritas Storage Foundation for Oracle RAC from Symantec within non global zones For more information on Storage Foundation for Oracle RAC support for non global zones please refer to the following README documentation https www secure symantec com connect sites default files sfrac_appnote_ zones 51sp1rp2_sol pdf It is assumed that the audience for this document holds prior knowledge of and experience with managing Veritas Cluster Server as well as Solaris zones Implementing Solaris Zones with Veritas Cluster Server from Symantec and Veritas Cluster file System High Availability 6 0 from Symantec INTERACTION BETWEEN VERITAS CLUSTER SERVER AND SOLARIS LOCAL ZONES In a Solaris environment the Veritas Cluster Server daemon High Availability Daemon HAD will always run within the context of global root zone of the Solaris 10 OS Traditionally Veritas Cluster Server has the capability to manage applications running in the global zone in what can be considered a cla
78. tly covered information regarding Solaris 11 will be provided only within the context of the ZFS appendix Further elaboration will be incorporated at a later date e With Veritas Storage Foundation 6 0 certain VxFS commands are supported within the context of a non global zone However for the purposes of this document that configuration option will not be covered INTRODUCTION With the release of Solaris 10 Sun Microsystems introduced the concept of a Local Zone Zones are an isolation mechanism for applications in which the application is executed within the confines of a zone From an application view the zone appears to be an independent system where the application gets exclusive use of system resources including processor and memory as well as access to specific file systems without risk of interfering with other applications From an implementation view a local zone does not instantiate a separate OS kernel as is done with Virtual Machines or para virtualization but rather zones operate as resource containers with independent user control and file systems In effect zones themselves are an instance of the user space portion of the OS Zones allow the system administrator to isolate an application and manage system resource allocation between applications running in other zones Zones extend the concepts of resource management from simply controlling resource allocations between applications to more robust isolation wher
79. umes are treated no differently than any other data source when configuring these types of snapshots The process begins with installing but not configuring a new zone root image This will act as the golden image source for all other zone roots and therefore only requires to be installed one time Create Source Zone Root File System Please refer to the section that previously covered creating disk groups volumes and file systems for use with zone roots Preparing a Golden Image Zone One time Operation 56 Implementing Solaris Zones with Veritas Cluster Server from Symantec and Veritas Cluster file System High Availability 6 0 from Symantec From Either Node gt zonecfg z gold_image_zone gold_image_zone No such zone configured Use create to begin configuring a new zone zonecfg gold_image_zone gt create b zonecfg gold_image_zone gt set zonepath zones gold_image_zone zonecfg gold_image_zone gt set autoboot false zonecfg gold_image_zone gt commit zonecfg gold_image_zone gt verify zonecfg gold_image_zone gt exit gt chmod 700 zones gold_image_zone gt zoneadm z gold_image_zone install gt zoneadm z gold_image_zone detach gt umount zones gold_image_zone Creating the Zone Root Snapshot Preparing Zone Root Volume for Snapshot Operations One Time Operation gt vxsnap g cfs_zroot_dg prepare cfs_zroot_vol Adding a Snapshot Mirror Assign a free disk for this operati
80. untPoint gt hares modify zoneroot_MNT MountPoint zones endzone base sys node2 gt hares modify zoneroot_MNT MountPoint zones calzone base sys node gt hares local zoneroot_MNT BlockDevice gt hares modify zoneroot_MNT BlockDevice endzone_zroot endzone_root sys node2 gt hares modify zoneroot_MNT BlockDevice calzone_zroot calzone_root sys node gt hares modify zoneroot_MNT FSType zfs gt hares modify zoneroot_MNT FsckOpt n gt hares modify zoneroot_MNT Enabled 1 38 Implementing Solaris Zones with Veritas Cluster Server from Symantec and Veritas Cluster file System High Availability 6 0 from Symantec gt hares online zoneroot_MNT sys node gt hares online zoneroot_MNT sys node2 gt chmod 700 zones calzone base amp from node gt chmod 700 zones endzone base amp from node2 Add NIC Resource From Either Node gt hares add zone_NIC NIC myzone_SG gt hares modify Zone NIC Device bgeO gt hares modify zone_NIC Critical 0 gt hares modify zone_NIC Enabled 1 Add IP Resource From Either Node gt hares add myzone_IP IP myzone_SG gt hares modify IP Critical 0 gt hares modify IP Device bgeO gt hares local IP Address gt hares modify IP Address 10 10 10 4 sys node gt hares modify IP Address 10 10 10 5 sys node2 gt hares modify IP NetMask 255 255 240 0 gt hares modify IP Enabled 1 Configure Resource Dep
81. ver configuration here would simply localize the Zone Name attribute in the Containerlnfo property for the service group to each respective global zone and thus only the application would failover This option does however require an independent zone for each physical host and as such you would be required to maintain user accounts application installs and DNS entries accordingly You can consider this scenario more analogous to traditional physical host clustering Application IP Address Configuration Solaris supports the ability to configure an application s virtual IP address within the zone configuration At first glance this appears to be an effective method of managing the virtual IP since the IP will be brought up and down along with the local zone However doing this has the following disadvantages e Since the IP isn t being monitored P related faults won t be detected It also worth noting that when cloning a zone root having the configuration as generic as possible makes for simpler provisioning Veritas Cluster Server supports local zones configured with or without the virtual IP but given the disadvantages mentioned above best practices dictate leaving the network information out of the zone configuration and using the IP agents in Veritas Cluster Server to control the virtual IP This topic is covered in more detail in the section Defining the Non Global Zone CONFIGURING A LOCAL ZONE TO WORK WITH CLUSTER FILE SYST
82. with Veritas Cluster Server from Symantec and Veritas Cluster file System High Availability 6 0 from Symantec From Both Cluster Nodes Define the Zone on each corresponding Physical Node Example From Node1 gt zonecfg z calzone create b set zonepath zones calzone base set autoboot false end commit verify exit gt zoneadm z calzone install gt zoneadm z calzone boot gt zlogin C calzone Follow Prompts for configuring zone Once complete you can use the keystroke to exit the console and return to the OS prompt gt cp etc resolv conf zones calzonezone base root etc gt cp etc nsswtich zones calzone base root etc This procedure will only work consistently for shared IP zones Configure the Second Zone on Node 2 From Node 2 gt zonecfg z endzone create b set zonepath zones endzone base set autoboot false end commit verify exit gt zoneadm z endzone install gt zoneadm z endzone boot gt zlogin C endzone Follow Prompts for configuring zone Once complete you can use the keystroke to exit the console and return to the OS prompt 48 Implementing Solaris Zones with Veritas Cluster Server from Symantec and Veritas Cluster file System High Availability 6 0 from Symantec gt cp etc resolv conf zones endzone base root etc gt cp etc nsswtich zones endzone base root etc This procedure will only work consistently for
83. you will need to export the zone configuration to each node gt zonecfg z newzone newzonel No such zone configured Use create to begin configuring a new zone zonecfg newzonel gt create b zonecfg newzone1 gt set zonepath zones newzone base zonecfg newzone1 gt set autoboot false zonecfg newzone1 gt commit zonecfg newzone1 gt verify zonecfg newzonel gt exit Export Zone Configuration to All Cluster Nodes From Node gt zonecfg z newzone export f tmp newzonezone cfg gt scp tmp newzone node2 tmp repeat for all nodes From All Remaining Nodes gt zonecfg z newzonezone f tmp newzonezone cfg From Node1 gt zoneadm z newzone attach F gt zoneadm z newzone boot gt zlogin C newzone a Follow the normal process for configuring Zone OS Adding the New Zone to Your Veritas Cluster Server Configuration Implementing Solaris Zones with Veritas Cluster Server from Symantec and Veritas Cluster file System High Availability 6 0 from Symantec From Node1 Please refer to the previous sections for the steps on how to add a new zone to a Veritas Cluster Server configuration 59 Implementing Solaris Zones with Veritas Cluster Server from Symantec and Veritas Cluster file System High Availability 6 0 from Symantec Appendix C Applying Patches to Systems with Zones Under Veritas Cluster Server Control Much of the discussion surrounding Solaris zones centers on the best practi

Symantec White Paper - Implementing Solaris Zones with Veritas

Contents

Download Pdf Manuals

Related Search

Related Contents