I have a NBU 701 installation consisting of the following:
Master Server - AIX 6.1
Media Server - AIX 6.1
Media Server - W2K8R2 (VMWare backup host)
All the three servers are fibre attached to an IBM library containing 4 x LTO5 drives.
SSO is operative wth the following settings:
Master Server - Scan Ability 0
Media Server AIX - Scan Ability 5
Media Server W2K8R2 - Scan Ability 5
When the environment is started (Master then Media's) all comes online - with the Media Servers becoming Active for Tape & Disk after approximately 7 minutes.
The AIX Master server resides on an AIX HACMP cluster - although NBU has not been installed in cluster mode. NBU was installed on one node, the resource group (containing /usr/openv filesystem) failed over to the second node, and NBU installed again. The NBU environment operates succesfully when running on either node. We do not want NBU to failover due to loss of services etc., the only time a failover will occur is due to a total server failure.
The only issue we have is following a failover.
The method we were using is as follows:
1. All servers up and running OK
2. We failover the Master Server to the other cluster node (different physical server - also has HBA's attaching to the tape library SAN)
3. After approximately 7 minutes the AIX and Wndows Media servers beome active for tape and disk
4. When we run jobs to backup through the Windows Media Server and also the AIX Media Server the Windows Media Server job runs OK but the AIX Media Server job sits there with a status of "WAITING FOR SHARED DRIVE SCAN TO STOP". This message does not go away - and the job on the Windows Media Server completes successfully.
If we perform the following all works OK:
1. Use the nbemmcmd command to set the SCAN ABILITY to 0 for the Master Server (which has just been failed over) - although it is already set to 0 ! As we have used te nbemmcmd to change SCAN ABILITY then we need to restart services EVERYWHERE......
2. Stop NBU services on both media servers
3. Stop NBU services on the Master
4. Start NBU Services on the Master
5. Start NBU services on both Media Servers
6. Wait for 7 minutes for both media servers to come active for tape and disk
7. Run backup jobs through both media servers - all works OK.
My questions are as follows:
1. Why are we having to re-enter the nbemmcmd to set the SCAN ABILITY for the master to 0, when it is already 0 ?
2. Where is the 7 minute wait coming from until the media servers come active for tape and disk ? I am guessing we are waiting for a retry interval.....
After spending a lot of time, sometimes going around in circles, on tis one - can anyone provide any suggestions ??
Thanks,
AJ.
