File Device Class Status Critial for no reason

[ivandem]

I'm running SP 7.1.7.200 on AIX. I've got three file device classes that are showing as "Critical" under status in the OC. The security is set as "drwxr-xr-x tsminst1 tsmsrvrs"

Is there something else I should be looking at?

TIV.
John

 

[marclant]

Hover over the critical symbol next to the device class for more info.

If that doesn't give enough information, check the activity log for messages regarding the devclass:
query actlog search={device_class_name} begindate=-7

 

[ivandem]

Hovering displayed nothing, and i tried everywhere on the each line. There is also nothing in the act log that give any clue.

Here is the output for actlog query for one of the devices..

• ANR2017I Administrator JOHND issued command: DEFINE DEVCLASS VMCTLDEV DEVTYPE=file MAXCAPACITY=10G DIRECTORY=/dcp1/vmctl (SESSION: 29992)
• ANR2203I Device class VMCTLDEV defined. (SESSION: 29992)
• ANR2017I Administrator JOHND issued command: DEFINE STGPOOL VMCTL VMCTLDEV POOLTYPE=PRIMARY MAXSCRATCH=100000000 DESCRIPTION='FIle pool for VM control files' (SESSION: 29971)
• ANR2200I Storage pool VMCTL defined (device class VMCTLDEV). (SESSION: 29971)
• ANR2017I Administrator JOHND issued command: QUERY DEVCLASS vmctldev F=d (SESSION: 31164)
• ANR2017I Administrator JOHND issued command: UPDATE DEVCLASS vmctldev DIR=/dcp1/vmctl,/dcp1b/vmctl (SESSION: 31163)
• ANR2205I Device class VMCTLDEV updated. (SESSION: 31163)

No errors; nothing to indicate an issue.

I should also mention that there is not problem accessing any of the directories. Nothing has been impacted, other then the seeming false Critical status

 

[marclant]

Might just be low on space then

 

[ivandem]

Not that either - plenty of space for each of them.

 

[marclant]

Click on Overview, then Alerts. There may be an alert with more info.

 

[DanGiles]

I am seeing a similar issue, but I am running db2 commands to try to clean up orphaned chunks. The storage pool is marked critical, but when you drill down to the directory level in the OC everything is green.

 

[DazRaz]

Lower your maxscratch down to the amount of disk space you have. I don't believe you have space for 100,000,000 x 10G virtual tapes. I believe it lets you over allocate by up to 30% but I suspect you are way over that.

 

[ivandem]

Thanks for the reply DazRaz. Your suggestion worked for the two device classes that had a storage pool associated with it. Who would have thought the IBM default would cause this.

The one remaining device class is used only for DB backups - this still remains as "Critical". It was out of space as the backup files were not being deleted. I have resolved the space issue this morning but it still has not changed. perhaps this just take some time?

 

[DazRaz]

I have one TSM server which uses a FILE device class for the DB backups. It ran out of space and we fixed by cleaning up old backups with "del vohist tod=-5 t=dbb" and scheduled this to run daily. Just had a look and it also is showing as critical however the filesystem is only at 59% full. I think it may be a bug.

 

[droach]

We had the same issue for a FILE device class that was only used for TSM's DB backups. The volume that was defined to the DEVC was a 1TB volume with 800GB free. So, plenty of space on the volume. The problem was that we/(I) had a the Device Class's setting 'MAXCAPacity' set to 100GB and the MOUNTLimit set to 20. Doing a little math 100GB x 20 = 2TB. Hmmm, no wonder Spectrum was complaining about a 50% threshold exceeded.

Kudos to my coworker for figuring this out and finding this article: http://www-01.ibm.com/support/docview.wss?uid=swg21696240