11-24-2014 04:20 AM
hi,
if i like to export an DG from an node,for an maintenace reason or when the VCS was not able to bring down the service group,first i need to offline the resource? hares -offline then vxdeport DG?
tnx a lot
Solved! Go to Solution.
11-24-2014 04:45 AM
If application is under VCS control, then so should diskgroup - to verify use:
hares -display -type DiskGroup -attribute DiskGroup
If this shows diskgroup, then when you offline group then diskgroup should be deported.
If VCS cannot bring down resources, then you would need to manually:
To check what service group diskgroup is in you can use:
hares -display -type DiskGroup | awk '$2 ~ /^Group$|^DiskGroup$/'
Mike
11-24-2014 05:22 AM
This seems like a theoretical question and not a real issue?
The steps to perform will entirely depend on the reason why diskgroup could not be offlined.
The reason will be logged in engine_a log.
The engine log will tell you exactly what is wrong and you will know and understand what to do.
Best to be pro-active and prepare properly for maintenance.
Especially disk/san/hardware maintenance.
Ensure everything in VCS and Storage Foundation is offline before you start with hardware maintenance.
Best to even shutdown servers.
As per Mike's excellent post - the diskgroup is normally the last of the resources to be offlined. Any other not going offline will prevent dg from being offlined.
The same as without cluster.
You can for example not deport a diskgroup while filesystems are mounted.
Filesystems cannot be unmounted while application is running (even someone cd'ed into the mountpoint).
It is almost impossible to prepare for every theoretical situation in forum discussions....
11-24-2014 04:45 AM
If application is under VCS control, then so should diskgroup - to verify use:
hares -display -type DiskGroup -attribute DiskGroup
If this shows diskgroup, then when you offline group then diskgroup should be deported.
If VCS cannot bring down resources, then you would need to manually:
To check what service group diskgroup is in you can use:
hares -display -type DiskGroup | awk '$2 ~ /^Group$|^DiskGroup$/'
Mike
11-24-2014 05:22 AM
This seems like a theoretical question and not a real issue?
The steps to perform will entirely depend on the reason why diskgroup could not be offlined.
The reason will be logged in engine_a log.
The engine log will tell you exactly what is wrong and you will know and understand what to do.
Best to be pro-active and prepare properly for maintenance.
Especially disk/san/hardware maintenance.
Ensure everything in VCS and Storage Foundation is offline before you start with hardware maintenance.
Best to even shutdown servers.
As per Mike's excellent post - the diskgroup is normally the last of the resources to be offlined. Any other not going offline will prevent dg from being offlined.
The same as without cluster.
You can for example not deport a diskgroup while filesystems are mounted.
Filesystems cannot be unmounted while application is running (even someone cd'ed into the mountpoint).
It is almost impossible to prepare for every theoretical situation in forum discussions....