Lan free Backup problem

[prabodh99in]

Hi,

I am taking a backup of File server using TSM 5.4
Total 700 GB data is there. But Full backup taking 4 days
to complete even with the Lan free backup.
Below are some details which i am getting at end of incremental Scheduled backup. Request you to help...

Total number of objects inspected: 963,154 (SESSION: 5415)
12/04/10 04:34:30 ANE4954I (Session: 5415, Node: MUPFS01)

Total number of objects backed up: 15,869 (SESSION: 5415)
12/04/10 04:34:30 ANE4958I (Session: 5415, Node: MUPFS01)
Total number of objects updated: 0 (SESSION: 5415)
12/04/10 04:34:30 ANE4960I (Session: 5415, Node: MUPFS01)
Total number of objects rebound: 0 (SESSION: 5415)
12/04/10 04:34:30 ANE4957I (Session: 5415, Node: MUPFS01)
Total number of objects deleted: 0 (SESSION: 5415)
12/04/10 04:34:30 ANE4970I (Session: 5415, Node: MUPFS01)
Total number of objects expired: 380 (SESSION: 5415)
12/04/10 04:34:30 ANE4959I (Session: 5415, Node: MUPFS01)
Total number of objects failed: 1 (SESSION: 5415)
12/04/10 04:34:30 ANE4965I (Session: 5415, Node: MUPFS01)
Total number of subfile objects: 0 (SESSION: 5415)
12/04/10 04:34:30 ANE4961I (Session: 5415, Node: MUPFS01)
Total number of bytes transferred: 32.75 GB (SESSION: 5415)
12/04/10 04:34:30 ANE4971I (Session: 5415, Node: MUPFS01) LanFree data bytes: 32.45 GB (SESSION: 5415)
12/04/10 04:34:30 ANE4963I (Session: 5415, Node: MUPFS01)
Data transfer time: 148.58 sec (SESSION: 5415)
12/04/10 04:34:30 ANE4966I (Session: 5415, Node: MUPFS01) Network data transfer rate: 231,139.96 KB/sec (SESSION: 5415)
12/04/10 04:34:30 ANE4967I (Session: 5415, Node: MUPFS01) Aggregate data transfer rate: 1,508.52 KB/sec (SESSION: 5415)
12/04/10 04:34:30 ANE4968I (Session: 5415, Node: MUPFS01) Objects compressed by: 0% (SESSION: 5415)
12/04/10 04:34:30 ANE4969I (Session: 5415, Node: MUPFS01) Subfile objects reduced by: 0% (SESSION: 5415)
12/04/10 04:34:30 ANE4964I (Session: 5415, Node: MUPFS01) Elapsed processing time: 06:19:26 (SESSION: 5415)
12/04/10 04:34:30 ANR0403I Session 5416 ended for node MUPFS01 (WinNT). (SESSION: 5416)
12/04/10 04:34:30 ANR2507I Schedule MUPFS01_INCR_BKP for domain FLATFILE started at 12/03/10 22:15:00 for node MUPFS01 completed
successfully at 12/04/10 04:34:30. (SESSION: 5415)

 

[GregE]

By the numbers of files involved, this data appears to be the type of which you should not be doing LANfree backup. Lots of small files going to directly to tape is going to be slow and LANfree is not meant for that type of data backup. You should change this to backup to disk and monitor the performance change.

Have you already tried a LAN backup to disk?

 

[prabodh99in]

No I have not tried that.
I will try the same and will get back to you.
thanks.

 

[prabodh99in]

Weekly file serrver backup is running. It will take three to complete.
Meanwhile i want to inform you one more thing related to problem that I observed session of Storage agent for library sharing is getting immediately closed once itt starts. After some time again task gets repeat. Is it related to File server backup probelm ?

 

[hogmaster]

You are right the session should not open and cloe like that, check yor mount retention in the device class Also check the client errorlog
and as GregE says this data should go to a diskstgpool in less than 2 hours using a 100 mb lan connection. What OS is the client running?
Have you checked the db cache hit ratio?

12/04/10 04:34:30 ANE4966I (Session: 5415, Node: MUPFS01) Network data transfer rate: 231,139.96 KB/sec (SESSION: 5415)
12/04/10 04:34:30 ANE4967I (Session: 5415, Node: MUPFS01) Aggregate data transfer rate: 1,508.52 KB/sec (SESSION: 5415)

 

[sgabriel62]

In my experience with file servers, you are spending countless hours with client/server mediation efforts for time stamp changes. Here is what I would due, taking into account your servers is only 900 thousand files - consider using Memory efficient backups and Journaling of selected "data" drives - place these into separate schedules so that you reduce the number of files being inspected at any one session. File servers should never to Lanfree - if anytime - Image based with incremental if you can swing it over. But Id wager that if you isolate your OS/APP drives from your data drives, you'll see an improvement.

Hope this helps

 

[prabodh99in]

mont retention details -
Mount Wait (min): 30
Mount Retention (min): 240

File server is running & storage agent running on windows server & TSM is on AIX.

 

[sgabriel62]

why would you wait this long - change both values to 2

 

[prabodh99in]

is it necessary to restart the TSM server after changing mentioned parameter in dsmserv.opt file ?

 

[hogmaster]

no, you do not need to restart the server , only the stgagent.
It is important that the vaules are identical amongst all the machines using the device.

 

[prabodh99in]

I have set both the values to 2 but then also facing same problem.

 

[prabodh99in]

I have tried LAN backup on Disk. Tell me one thing, IS the Tivoli SANergy is required for the SAN backup on Disk.

 

[GregE]

Yes SANergy is required. Most people will tell you it's obsolete and not widely implemented anymore. Search the forum for it.

 

[chad_small]

LAN-Free to disk should not make that much of a difference. The problem has to be something else because I've seen servers do millions of files in a 5-6 hours. What I would check for is a file system where one directory has the majority of the files in it. I've seen apps that dump small 1K files in a directory and eventually TSM and the OS are brought to their knees trying to list the contents. Also how much memory do you have on the server? What OS level? How many CPU's?

 

[sgabriel62]

I have taken a closer look at the beginning summary of this thread - and I am not seeing a 4 day backup routine but a 6 hr backup routine where only 32 Gb went by in those 6 hrs. I'm also seeing network thruput of 231k/sec - that is what Half Duplex speeds - since it seems your configurations are now becoming more in tune - lets take a look at the networking side - the switch settings- the tape drives - etc...

 

[prabodh99in]

How to place Memory efficient backups and Journaling of selected "data" drives in different schedules?