Scribd
Upload
3 views

SFHA Upgrade Procedure

Uploaded by

fast2win2029
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as DOCX, PDF, TXT or read online on Scribd
3 views

SFHA Upgrade Procedure

Uploaded by

fast2win2029
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as DOCX, PDF, TXT or read online on Scribd
You are on page 1/ 2

- check size of /var/tmp and /opt vol (/var/tmp 5G free min if you

download iso image also, /opt 1G)


- check for existing vxfendg if fencing is enabled!!!!
if missing, some nodes crash!!! go for add new
- analyze system(s) sort, VeritasScript.sh => upload .xml data files to
sort portal, print analyze as .pdf
- save actual cluster configuration (e.g. /var/tmp/linux_patch/backup)

- check fencing
- evacuate service groups from the node

- reboot node in front of any activities


- freeze all nodes persistent
- stop vcs, vxodm, vxfen, gab & llt
- mv /etc/llttab to /etc/llttab.<date> to prefent any cluster
fuctionality

- on failover cluster: stop vcs, vxodm, vxfen, gab, llt on all nodes
- on CVM / CFS: stop vcs and fencing on all nodes, vxodm, gab & llt on
patching node

- check versions of all dg's, volumes and protocolversion on all nodes


- let the node in runlevel 3
- on failover: start llt and gab by hand
- extract & install VRTS_SF_HA_Solutions_6.0.1....
- extract & install sfha..rhelXX..6.0.3
- extract & install vm...6.0.3.. hot-fix
- extract & install VRTSaslapm ...
- extract & install all other hot-fix archives

- mv /etc/llttab to /etc/llttab.<date>
- reboot
- take care on boot messages
- mv llttab back, start llt by hand
- start gab by hand & check
- start fencing by hand & check
- start vxodm
- start vcs & check status
- stop vcs, vxodm, vxfen, gab, llt

- check all fs at least of version 7!!!, bring it up in single steps if


needed
- start llt, gab, vxfen, vxodm vcs by hand on service holder node
- unfreeze service holder node
- stop all application, services, db's
- bring down all recources by stopping vcs regular
- stop vxodm, vxfen, gab & llt
- start llt, gab, vxfen, vxodm vcs by hand on patching node
- unfreeze node
- bring service groups online on this node
- leave unpatched node untouched until next phase !!!!

NEXT PHASE
- reboot other node
- do all patch activities like first node
- reboot node and let it join to the cluster
- switch ALL service groups on the second node (and back again, if
needed)
- distribute service groups like state before

- upgrade all filesystems to version 9


- on CVM / CFS upgrade protocolversion to 120 first
- upgrade dg's to 180

You might also like