Video Screencast Help

The backup that takes forever!

Created: 09 Sep 2013 | 5 comments

I have a backup policy that never seems to finish and destroys my window. Anyone see anything suspicious... i cant figure out why it takes sooo long or what the deal is.???

We run NBU This policy is a full, ALL_LOCAL_DRIVES, scheduled for off-hours every day, retention for 2 weeks. its a 2003 sp2 32bit VM server. 1 local drives, total size - 80GB.

Job Overview - started at 1:20:48 AM, Elapsed: 10:27:52 30% complete, 1466 minutes remaining.

kilobytes written: 18400000, Current Files Written: 2805500, Estimated Kilobytes: 60387436, Estimated Files: 5965220

File List on Job:

SET SNAP_ID=smsserver_1378704016
BACKUP C:\ USING \\?\GLOBALROOT\Device\HarddiskVolumeShadowCopy313\ OPTIONS:ALT_PATH_PREFIX=C:\Program Files\VERITAS\NetBackup\temp\_vrts_frzn_img_1304_1,FITYPE=MIRROR,MNTPOINT=C:\,FSTYPE=NTFS

Detail Status:

9/9/2013 1:20:48 AM - Info nbjm(pid=2592) starting backup job (jobid=360643) for client smsserver, policy Windows_2003Servers, schedule Full 
9/9/2013 1:20:49 AM - estimated 60387436 Kbytes needed
9/9/2013 1:20:49 AM - Info nbjm(pid=2592) started backup (backupid=smsserver_1378704048) job for client smsserver, policy Windows_2003Servers, schedule Full on storage unit horus-hcart-robot-tld-1
9/9/2013 1:20:49 AM - started process bpbrm (6852)
9/9/2013 1:20:55 AM - Info bpbrm(pid=6852) smsserver is the host to backup data from    
9/9/2013 1:20:55 AM - Info bpbrm(pid=6852) reading file list from client       
9/9/2013 1:20:58 AM - connecting
9/9/2013 1:21:02 AM - Info bpbrm(pid=6852) starting bpbkar32 on client        
9/9/2013 1:21:02 AM - connected; connect time: 00:00:04
9/9/2013 1:21:10 AM - Info bpbkar32(pid=2296) Backup started          
9/9/2013 1:21:10 AM - Info bptm(pid=3404) start           
9/9/2013 1:21:10 AM - Info bptm(pid=3404) using 65536 data buffer size       
9/9/2013 1:21:10 AM - Info bptm(pid=3404) setting receive network buffer to 263168 bytes     
9/9/2013 1:21:10 AM - Info bptm(pid=3404) using 30 data buffers        
9/9/2013 1:21:11 AM - Info bptm(pid=3404) start backup          
9/9/2013 1:21:11 AM - Info bptm(pid=3404) backup child process is pid 9500.9624      
9/9/2013 1:21:11 AM - Info bptm(pid=3404) media id 340066 mounted on drive index 6, drivepath {2,0,2,0}, drivename IBM.ULTRIUM-TD4.002, copy 1
9/9/2013 1:21:11 AM - Info bptm(pid=9500) start           
9/9/2013 1:21:11 AM - mounted
9/9/2013 1:21:11 AM - positioning 340066 to file 84
9/9/2013 1:21:13 AM - positioned 340066; position time: 00:00:02
9/9/2013 1:21:13 AM - begin writing
9/9/2013 1:21:18 AM - Info bpbkar32(pid=2296) change journal NOT enabled for <C:\>

Operating Systems:

Comments 5 CommentsJump to latest comment

Nate.D's picture

Do you have any other similar machines that dont take that long to backup? Is there anything special networking wise about this machine? I would think to look there first if you are not getting any errors out of netbackup.

If I was helpful in solving your issue please mark my post with a thumbs up or a solution!  Have a great day :)

quebek's picture

Few things to check:

  1. network settings on client end (100Mb vs 1Gb; full vs half duplex)
  2. disk fragmentation -> if huge try to defragement
  3. block size on the client filesystem (most probably 4kB -> not good)
  4. if biggest disk is a raid array - check its stripe size and if not the biggest possible look at point 6
  5. is system patched to latest available level
  6. points 2,3 and 4 do in one step: if fragementation is huge and block size is 4kB (and average file size is big enough) backup (make sure the backup is 100% firm/restorabel), recreate the LUN with biggest possible stripe size, reformat partition with 16 or 32 or 64 kB block size (depending on your average file size) - restore data.
  7. change block size to 256kB -> SIZE_DATA_BUFFERS
  8. create bpbkar log on client end to 'see more'

This is what I would do - of course....

watsons's picture

a VM server, I assume it's with Netbackup client installed (please clarify), 80GB local drive, can you check how many files it has?

Also look like a MS-Windows policy, or is it Windows-Flashbackup? 

A screenshot of the policy attributes, or the output of "bppllist -L <policyname>" would be helpful

revarooo's picture

You're backing up over 2 million files. That's a fair few files. Can you run:

bpbkar -nocont <fileselection>   > NUL

How long does that take to complete?

Mark_Solutions's picture

I am with Revaroo here - estimated over 6 million files

This client needs its own policy and where ever all that data it it needs splitting into lots of streams to get some decent throughout out of it.

As it is virtual have you tried backing it up via a VMware backup - that may be a better solution but the file listing may still cause you issues.

Worth a try as a VMware backup but if that doesn't work for you then find the data causing the issue, split it into lots of streams and try it that way.

Hope this helps

Authorised Symantec Consultant

Don't forget to "Mark as Solution" if someones advice has solved your issue - and please bring back the Thumbs Up!!.