No storage for client transaction

The customer wants to know if it is possible to get a request for improvement started so that TSM B-A clients will handle this problem file more elegantly, without blowing up server storage?

We added a few hundred Unix servers to ECS (object storage/S3) storage last night. These clients had V1 dedupe, so we just redirected them to the V2 dedupe. Having approximately 9TB for stgpooldirs, we only consumed a total of around 1TB last night. And yet two separate clients complained of a lack of storage in ECSPOOL. The likely culprit that I know of is the Linux sparse file "/var/log/lastlog", which can be ~500GB. The nature of TSM backup is to pre-allocate enough space for this file, even if it only condenses to maybe a few hundred MB.

Is it possible to get a request for improvement started so that TSM B-A clients will handle this problem file more elegantly, without blowing up server storage?

Team,


There are two warnings last night occurred around 9:00 pm, on the following clients

Two nodes have reported on this warning, SV65664(WinNT) on WINTEL2100
SV67477( Linux X86-64) on Unix2100

Both client schedules at end reported completed

Scheduled Start Actual Start Schedule Name Node Name Status
-------------------- -------------------- ------------- ------------- ---------
06/25/19 21:00:00 06/25/19 21:05:04 WINTEL2100 SV65664 Completed
06/25/19 21:00:00 06/25/19 21:04:35 UNIX2100 SV67477 Completed

For now, no issue, in future, tuning on client schedule may require.