Spectrum Protect DB file system space utilization

[rowl]

I am seeing something and was wondering if anyone knows how to look deeper into this. I have some Spectrum Protect databases around 3TB in size, but the DB file systems are showing 3.7TB of used space. If the instance is restarted this used space drops back down to 3TB, but over time climbs up again. It has been hinted at that this is due to some background database activity using these file systems as temp space. Is there any way to see what that activity is, how far along it is, and when it will return space to the file systems? I assume something is running that needs to finish, so I don't want to interrupt it.

Other members of my team are panicing and restarting the instance when they see this high file system space utilization and I would like to be able to find a view into this to see what is actually going on.

Thanks
-Rowl

 

[marclant]

It's true that database activity will use temporary space.

Some of it is your typical processes you see running in Q PROC. And then there's online reorgs that also occur in the background. Client sessions also causes database transactions.

I would not restart the instance to release the space, that's a waste of time and resources. I'd increase the database size if it's running too close to full.

 

[booman55]

If you login to DB2 as instance owner and connect to tsmdb1 you can do "db2 list utilities" to show DB2 processes running.

 

[Silapen]

Have you checked the size of active/archive logs?

 

[Rigido]

Have you checked the size of active/archive logs?

Do you mean this?

tsm: TSM_P7>q log f=d

           Active Log Directory: /tsm_p7/log01
                Total Space(MB): 16.384
                 Used Space(MB): 205
                 Free Space(MB): 16.179
Total Space of File System (MB): 20.480,00
 Used Space on File System (MB): 16.387,80
 Free Space on File System (MB): 4.092,20
          Archive Log Directory: /tsm_p7/arch
Total Space of File System (MB): 66.560,00
 Used Space on File System (MB): 9.644,60
 Free Space on File System (MB): 56.915,40
         Archive Log Compressed: No
           Mirror Log Directory:
Total Space of File System (MB):
 Used Space on File System (MB):
 Free Space on File System (MB):
 Archive Failover Log Directory: /tsm_p7/fail
Total Space of File System (MB): 10.240,00
 Used Space on File System (MB): 1.025,91
 Free Space on File System (MB): 9.214,09

 

[Rigido]

Sorry, I read "How" instead of "Have"

 

[Silapen]

Do you mean this?

tsm: TSM_P7>q log f=d

           Active Log Directory: /tsm_p7/log01
                Total Space(MB): 16.384
                 Used Space(MB): 205
                 Free Space(MB): 16.179
Total Space of File System (MB): 20.480,00
Used Space on File System (MB): 16.387,80
Free Space on File System (MB): 4.092,20
          Archive Log Directory: /tsm_p7/arch
Total Space of File System (MB): 66.560,00
Used Space on File System (MB): 9.644,60
Free Space on File System (MB): 56.915,40
         Archive Log Compressed: No
           Mirror Log Directory:
Total Space of File System (MB):
Used Space on File System (MB):
Free Space on File System (MB):
Archive Failover Log Directory: /tsm_p7/fail
Total Space of File System (MB): 10.240,00
Used Space on File System (MB): 1.025,91
Free Space on File System (MB): 9.214,09

If regular full DB backup is not performed, it can cause the archive log space grow. As I can see, this is not the case, since filespace is 66GB, and you mentioned 0.7TB.

 

[rowl]

Total Space of File System (MB): 4,022,627
Space Used on File System(MB): 3,846,333
Space Used by Database(MB): 3,147,076
Free Space Available (MB): 176,294

that is what I am seeing, the DB file systems have several 100GB of space consumed by something other than the DB itself.

I see that db2 list utilities shows RUNSTATS is currently running in TSMDB1.SD_CHUNK_LOCATIONS. Curious thing, that process started 15 minutes ago, but this space consumption has been going on for a day.

 

[Silapen]

Could you run this in TSM console "q act begindate=today-5 endd=today endtime=now s=reorg "

 

[rowl]

I see several reorgs logged, but all have start end end like this...

05/30/2019 03:00:52 ANR0317I Reorganization of indices for table SD_FRAGMENTED_CONTAINERS started.
05/30/2019 03:00:52 ANR0318I Reorganization of indices for table SD_FRAGMENTED_CONTAINERS ended with sqlCode 0.

q db f=d matches up on this time
Last Database Reorganization: 05/30/2019 03:00:52