From: Irvine Siggle (irvsiggle@tokuko.de)
Date: Wed Jan 02 2008 - 06:07:07 EST
Hello,
I know that this message might be somewhat inappropriate as this is only
related to Sun indirectly, but I need to finish this task urgent and have
no better address to ask for support. Sorry for that.
I have an old Veritas Cluster Server 2.4 running on SunOS 5.8. The Cluster
manages an Oracle instance, VxVM diskgroups & volumes and NFS shares.
My task is to add a new volume, mount it and add an NFS share for it.
Sounds simple, problem is that it is supposed to happen without a downtime
as the system is in constant usage. My fear is that while updating the VCS
configuration the cluster will be either unable to online the resource
because it's already up on the system, or unable to see it online because I
have not brought it up already and therefor try to switch the resource
group to the other node or produce some other service interruption.
Details:
- Diskgroup (vaultdg) is online _while_ doing the change:
#Resource Attribute System Value
vaultdg Group global vault_SG
vaultdg Type global DiskGroup
vaultdg AutoStart global 1
vaultdg Critical global 1
vaultdg Enabled global 1
vaultdg LastOnline global vn2
vaultdg MonitorOnly global 0
vaultdg ResourceOwner global unknown
vaultdg TriggerEvent global 0
vaultdg ArgListValues vn1 vaultdg 1 1 0
vaultdg ArgListValues vn2 vaultdg 1 1 0
vaultdg ConfidenceLevel vn1 0
vaultdg ConfidenceLevel vn2 100
vaultdg Flags vn1
vaultdg Flags vn2
vaultdg IState vn1 not waiting
vaultdg IState vn2 not waiting
vaultdg Probed vn1 1
vaultdg Probed vn2 1
vaultdg Start vn1 0
vaultdg Start vn2 1
vaultdg State vn1 OFFLINE
vaultdg State vn2 ONLINE
vaultdg DiskGroup global vaultdg
vaultdg StartVolumes global 1
vaultdg StopVolumes global 1
1. On node vn2 I create the volume using Veritas Storage Administrator GUI
and create the filesystem
2. On both nodes I create the mountpoint
3. On vn2:
hares -add NewVolume Volume vault_SG
hares -modify NewVolume ... (AutoStart=1, Critical=1, DiskGroup=vaultdg,
Volume=NewVolume)
hares -add NewVolume_mount Mount vault_SG
hares -modify NewVolume_mount ... (AutoStart=1, Critical=1,
BlockDevice=/dev/vx/dsk/vaultdg/NewVolume, FSType=vxfs, MountOpt=logging,
MountPoint=/vault/NewVolume)
hares -add NewVolume_share Share vault_SG
hares -modify NewVolume_share ... (AutoStart=1, Critical=0,
PathName=/vault/NewVolume)
The last things left to do is mounting the volume on node vn2, creating the
NFS share and then - this is the part that worries me - set the Enabled
Attributes to 1 in the cluster config.
Now back to my main question: Is this the correct way to add the resources
without a downtime and will it work this way? Or do I have to fix my
process?
Thanks for reading this far and for any possible help. I will post a
summary to the list when I have finished this.
Irvine
_______________________________________________
sunmanagers mailing list
sunmanagers@sunmanagers.org
http://www.sunmanagers.org/mailman/listinfo/sunmanagers
This archive was generated by hypermail 2.1.7 : Wed Apr 09 2008 - 23:42:38 EDT