How much is a client backing up?

[randyman]

Greetings,

I'm running TSM Server 5.5.2.1 on Windows 2003 server. 99% Windows clients. I have a 6.2.1 Client running on Windows that has been backing up way under 100GB a night for the longest time. About a week ago, its now backing up nearly 700GB. Needless to say, this has dramatically affected the backup times! So, the server owner does not know why, and I can't seem to figure out why. Nothing has changed. Now, I'm sure if I take the server owner and/or application owner to a dark room tie him to a chair and shine a bright light in his face, I might get an answer.

But, since thats not going to happen, is there a way to see what specifically is being backed up while a session is running? What file specifically is it working on? Or someway to see what files are backed up on a particular node on any given night? etc...

Thanks!

 

[moon-buddy]

Can't you look at the node's backup logs (dsmsched.log) to see what changed as compared when it was just backup up 100GB?

 

[randyman]

I'm not sure what you're meaning. The sched log just has the folder...no sizes...for example..these are mountpoints:
12/10/2010 01:00:59 Successful incremental backup of '\\casprdc\e$\rpt-tlog'

12/10/2010 01:00:59 Successful incremental backup of '\\casprdc\e$\rpt-tempdb'

12/10/2010 01:00:59 Successful incremental backup of '\\casprdc\e$\rpt-data04'

12/10/2010 01:00:59 Successful incremental backup of '\\casprdc\e$\rpt-data03'

12/10/2010 01:00:59 Successful incremental backup of '\\casprdc\e$\rpt-data02'

12/10/2010 01:01:00 Successful incremental backup of '\\casprdc\e$\rpt-data01'

I'm interested in to see what this 700GB actually consists of that its backing up each night. Here's the end of the sched log

12/10/2010 09:10:44 --- SCHEDULEREC STATUS BEGIN
12/10/2010 09:10:44 Total number of objects inspected: 362
12/10/2010 09:10:44 Total number of objects backed up: 18
12/10/2010 09:10:44 Total number of objects updated: 0
12/10/2010 09:10:44 Total number of objects rebound: 0
12/10/2010 09:10:44 Total number of objects deleted: 0
12/10/2010 09:10:44 Total number of objects expired: 8
12/10/2010 09:10:44 Total number of objects failed: 0
12/10/2010 09:10:44 Total number of subfile objects: 0
12/10/2010 09:10:44 Total number of bytes transferred: 698.02 GB
12/10/2010 09:10:44 Data transfer time: 16,650.64 sec
12/10/2010 09:10:44 Network data transfer rate: 43,958.40 KB/sec
12/10/2010 09:10:44 Aggregate data transfer rate: 24,898.68 KB/sec
12/10/2010 09:10:44 Objects compressed by: 0%
12/10/2010 09:10:44 Subfile objects reduced by: 0%
12/10/2010 09:10:44 Elapsed processing time: 08:09:56
12/10/2010 09:10:44 --- SCHEDULEREC STATUS END

 

[fuzzballer]

This appears to be a SQL server. Did the DBAs do a large data dump or are they running some humongus queries/updates/maintenance jobs during your backup period?

 

[ypcat]

If you want to know exactly what is being backed up you need to switch your client side schedule log DSMSCHED.LOG to verbose mode by commenting out the QUIET option. As this will create a massive log with information about every file backed up we typically only do this for diagnostic purposes, and very quickly switch it back to QUIET. This should answer your question.

 

[randyman]

If I take them into a dark room, tie them to a chair and shine a bright light at them...then maybe...but they say not. I only see 1 .bak file and its no where near 700GB.

 

[fuzzballer]

Also in the schedlog there are sizes listed in bytes. You have to scroll through the days backup. Tedious

 

[randyman]

Thanks! I'll turn Quiet off next week when I can watch it and find the culprit. Good thinking.