cancel
Showing results for 
Search instead for 
Did you mean: 

Duplication jobs don't start

IanB
Level 4

We have NetBackup 6.5.2 on Windows 2003 Server Master/media server, backing up shares on a SAN (EMC CX3-40c) from clustered WindowsUDSS clients (4 Dell NX1950's), onto other LUNs on the SAN.  We have setup basic disk staging, to duplicate the images from the SAN backup LUNs (3 storage units) to our twin tape libraries.  The staging policy checks every 4 hrs to see if new images need duplicating.  Backup to disk is working fine, and 50% of the time the duplications work, but often they will only work for 1 of the 3 storage units (can be any one, it isn't consistent).  For the other units, the DSSU jobs just don't appear.  To get them started again we have to run BPDOWN and then BPUP, and hey presto! the missing DSSU jobs start and duplication works again - for a couple of days.

Any ideas?

Thanks in advance

Ian

11 REPLIES 11

jim_dalton
Level 6

This will be scant consolation Ian, but I've got issues with netb 6.5.2 (Solaris) skipping backups...some just dont run: failure I can acccept but if they never even fire off, then its a problem. Symantec support are investigating and have been for 3 weeks or so.

Tim

PatrikJ
Level 3
Certified

Hi Tim,

 

Have You got any solution on this issue? I have got exactly the same problem on Solaris.

It's really nasty when the duplication don't even start....

 

/Patrik

IanB
Level 4

Tim/Patrik,

Thanks for your messages.  The problem still occurs.  I had to BPDOWN/BPUP on Friday 12th Sept to wake up the duplications, but it's been OK since then (fingers crossed).  Our Master server is on Windows 2003 Server.

Ian

PatrikJ
Level 3
Certified

Hi Ian,

 

When this problem occures I just start a manual relocation and then it might work for some days.

I don't have to restart the nbu services.

 

/Patrik

PatrikJ
Level 3
Certified

HI,

 

I have made some more tests and here is my results.

 

First I verified that all stagings from the DSSU where in the schedule. nbpemreq -predict_all ....

Then I made a change to one of the DSSU's, No of concurrent jobs, for example.

Then when I ran nbpemreq -predict_all, all but the DSSU I changed was gone...

 

I got information from a NBU engineer at Symantec who is working on another case for me, that there is a issue with Etrack no ET1379870. I have now registerad a case on this. 

 

/Patrik

IanB
Level 4

Nice work, Patrik!

I checked a bit more - if you run the Admin Console but don't make a change to the DSSU, and OK out, the other scheduled DSSU jobs still disappear from nbpemreq, just leaving the one you looked at! 

You just have to restart the NetBackup Policy Execution Manager service, and hey presto! they're all back again.

Thanks for your help.

Ian

jim_dalton
Level 6

Just to keep running with this, Ive installed an engineering version of nbpem. The day following the patch, my problem remained: my conclusion - the patch didnt fix my problem.

Then for a few days everything ran as expected. Then one day nothing ran at all even via cron.

Following a restart of nbpem and a couple of other daemons everything has worked as normal. This I consider to be a different problem to my original problem of just a few occasional backups missing.

This I dont feel to be a fixed problem - if after patching the problem persists ( as it did ) then the problem hasnt been fixed and no argument. 

 

I wouldnt rely too much on what nbpemreq tells you: for a start the man page doesnt match the actual list of arguments but more worryingly its just wrong:

 

nbpemreq -predict_all -client_filter sap-node-a -date 09/23/2008 17:00:00

 

Predicted worklist including all jobs due to run between now and Tue Sep 23 17:00:00 2008

  00 sap-node-a           Daily_SAP_Online_A SAP_Daily            0          Mon Sep 22 15:00:00 2008

 

so its saying it'll run on the 22nd (correct) but it should also run on the 23rd.

Once the job has run on the 22nd, if you issue the command again , then it will tell you  it will run on the 23rd. So either Ive got the wrong end of the stick , or predict_all is in fact just predict (it does say "all jobs" !), in which case  whats going on!

 

My case is still open with symantec...

 

Tim 


 

Omar_Villa
Level 6
Employee

nbpem was totaly reengineered on 6.5.2 thats why it came up with a lot of issues upgrade to 6.5.2a and ask for the nbpem bin to Symantec, this fix a lot of schedule problems with none standard backups has Oracle, MSSQL and the Vault.

 

regards

PatrikJ
Level 3
Certified

Hi Omar,

 

I have 6.5.2A today and I have opened a case at Symantec, also referenced Etrack ET1379870, that should have something to do with this issue. But the engineer couldn't find the Etrack. He was going to investigate futher...

 

/PAtrik

Dave_Drummond
Level 3
I've also got a case open for this issue and already had the updated nbpem.exe  I received an updated bpdbm.exe which fixes the problem on my development system but makes it worse on my production system.  If you look at a DSSU and click okay then everything looks okay (checked using nbpemreq -due) until you realise that nothing is running at all !  nbpemreq -due shows active jobs but they do not appear on activity monitor.  A restart of nbpem recovers the situation.

PatrikJ
Level 3
Certified

A update:

 

This case is not solved yet. The etrack is a windows based issue. But I'm running Solaris 10.

 

To me it sounds as Solaris has a similar issue as windows.

 

/Patrik