scheduling problem

[storageadmin]

Hi..

I am using IBM Tivoli storage Manager at two of our offices. We r using version 5.3.0., on linux platform. and we use webadmin for all tasks. My problem is that whenever i schedule an incremental backup, it says backup completed within 1 min and no.of. objects inspected are 8-11 objects. If i manually backup the same data..it takes nearly 1 hour and no.of objects inspected are 3 to 4 lakhs. I am facing this problem only at one site. At other site scheduling is running fine.

Please advice me...

Regards
Storageadmin

 

[moon-buddy]

How do you start the TSM backup schedule? What OS is the problematic node on? Are you using prompted or polled mode? Can you post your dsm.opt/dsm.sys file as the case maybe?

 

[storageadmin]

Hi,

given below is the dsm.sys file. the OS we r using is Redhat Linux.

SErvername server_a
COMMMethod TCPip
TCPPort 1500
TCPServeraddress ipaddress
passwordaccess generate
txnbytelimit 2097152
resourceutilization 7

regards
storageadmin

 

[moon-buddy]

Storageadmin,

You have not given enough information:

How do you start the backup - through the scheduler?
Polling or prompted? I am guessing it is polling (default). If it is, do you have the dsmcagent running?

 

[storageadmin]

hi

Schedule mode ia set to ANY.

i define the schedule from webadmin and run dsmc sched thru command prompt from the server.

regards
storgageadmin

 

[max]

hi,
check into the web interface what kind of "object" and "option" you set into the schedule. You may be doing a schedule without the option "-subdir=yes", or you may have written the wrong objects.
You should as well check into the dsmsched.log what it is saving
hope this helps
max

 

[storageadmin]

hi..

TSM server and client are in same linux server.

 

[max]

ok, that doesn't matter, because client and server are two separate object ...

That means that when you run the scheduled process, it picks up the command from what you scheduled on the Server.
When you run the process manually, it is really the client to decide what and how to backup
cheers
max

 

[storageadmin]

HI..

Given below is the dsmsched.log

06/25/2007 15:38:42 --- SCHEDULEREC QUERY BEGIN
06/25/2007 15:38:42 --- SCHEDULEREC QUERY END
06/25/2007 15:38:42 Next operation scheduled:
06/25/2007 15:38:42 ------------------------------------------------------------
06/25/2007 15:38:42 Schedule Name: INCR_BKP
06/25/2007 15:38:42 Action: Incremental
06/25/2007 15:38:42 Objects: /vol01/*
06/25/2007 15:38:42 Options:
06/25/2007 15:38:42 Server Window Start: 15:50:59 on 06/25/2007
06/25/2007 15:38:42 ------------------------------------------------------------
06/25/2007 15:38:42 Command will be executed in 13 minutes.
06/25/2007 15:51:42
Executing scheduled command now.
06/25/2007 15:51:42 Node Name:SERVER1
06/25/2007 15:51:42 Session established with server UNIT1: Linux/i386
06/25/2007 15:51:42 Server Version 5, Release 3, Level 0.0
06/25/2007 15:51:42 Server date/time: 06/25/2007 15:51:42 Last access: 06/25/2007 15:38:42

06/25/2007 15:51:42 --- SCHEDULEREC OBJECT BEGIN INCR_BKP 06/25/2007 15:50:59
06/25/2007 15:51:42 Incremental backup of volume '/vol01/*'
06/25/2007 15:51:43 Successful incremental backup of '/vol01/*'

06/25/2007 15:51:45 --- SCHEDULEREC STATUS BEGIN
06/25/2007 15:51:45 Total number of objects inspected: 11
06/25/2007 15:51:45 Total number of objects backed up: 0
06/25/2007 15:51:45 Total number of objects updated: 0
06/25/2007 15:51:45 Total number of objects rebound: 0
06/25/2007 15:51:45 Total number of objects deleted: 0
06/25/2007 15:51:45 Total number of objects expired: 0
06/25/2007 15:51:45 Total number of objects failed: 0
06/25/2007 15:51:45 Total number of bytes transferred: 0 B
06/25/2007 15:51:45 Data transfer time: 0.00 sec
06/25/2007 15:51:45 Network data transfer rate: 0.00 KB/sec
06/25/2007 15:51:45 Aggregate data transfer rate: 0.00 KB/sec
06/25/2007 15:51:45 Objects compressed by: 0%
06/25/2007 15:51:45 Elapsed processing time: 00:00:03
06/25/2007 15:51:45 --- SCHEDULEREC STATUS END
06/25/2007 15:51:45 --- SCHEDULEREC OBJECT END INCR_BKP 06/25/2007 15:50:59
06/25/2007 15:51:45 Scheduled event 'INCR_BKP' completed successfully.
06/25/2007 15:51:45 Sending results for scheduled event 'INCR_BKP'.
06/25/2007 15:51:45 Results sent to server for scheduled event 'INCR_BKP'.

Regards,
storageadmin

 

[moon-buddy]

I believe Max has answered your question: -subdir=yes is missing.

 

[max]

yep,
you must have missed the option subdir=yes into the sched
cheers
max

 

[storageadmin]

hi max and moon buddy..

its working..thanku very much.

one more clarification..what i have observed is when i manually start backup..it takes 2 to 3 hrs to complete..but scheduled backup takes..nearly 4 hrs to backup same data.
what would be the reason for this.

regards,
storageadmin.

 

[BBB]

If your scheduled backup happens at the same time as a lot of other backups or something else that uses the network or san a lot then it will run slower because your TSM server (and possibly SAN) are busier.

 

[karunakamulu]

hi max,
this is karuna.im new to tsm .could u plz help me how to practice on commandline

 

[storageadmin]

hi,

we run daily backup only twice a day. there r no other backup's happening on tsm..the whole day. i have scheduled this backup..twice daily. if i start backup manually it takes less time compared to the scheduled backup.

 

[webxpertin]

You need to check the performance tuning parameters at the server - dsmserv.opt file. check the common paramateres which are also maintained in dsm.opt (example TCPWindowSize).

This issue can be fix with the help of proper performance tuning from the tsm server side.

 

[jharris]

It's possible that when you are manually backing up, you may be enabling client compression, but haven't specified the switch in your scheduled job.

Also check the number of objects inspected for both types of backups (manual/scheduled) to ensure that you're still comparing similar backup selections.