10-11-2017 10:28 PM
NBU 7.7.3
so i have this several SLP in queue for days now (2nd tape drive went south), and today i find those were in error 12. upon checking the details i get this error:
10/11/2017 13:18:11 - Error bpduplicate (pid=4241) cannot open /usr/openv/var/global/nbstserv/bid_file_rs_1507267508_0_0, No such file or directory
so in i went at the OS level and indeed the file is missing. my question is why?
the same error 12 i now have for those in queue duplication jobs that are now showing in red.
how do i find out why this duplication jobs disappeared? is there some kind of housekeeping that i'm not aware of? or was it intentional?
appreciate any help.
regards,
10-11-2017 11:15 PM
OMG! the housekeeping deleted it! https://www.veritas.com/support/en_US/article.000100674
why aren't this settings or defaults documented in an easy to find document?
10-11-2017 11:48 PM
Hi Rhino19ny,
I found this link which I think may relate to your problem:
NetBackup duplications fail with status 12: https://www.veritas.com/support/en_US/article.000100674
I'm guessing the BID file has been cleaned up, as your jobs have been queued up for a while. I hope the setting is still relevant for 7.7.3.
Hope this helps,
Steve
10-12-2017 12:19 AM
10-12-2017 12:48 AM - edited 10-12-2017 12:51 AM
Hi Rhino,
When I originally looked, I thought that technote was for 7.6.x. Have rechecked and it is for 7.7 as well
There was another link on VOX from 2015 that Marianne was involved in, so she may have some extra comments:
https://vox.veritas.com/t5/NetBackup/NetBackup-7-6-0-1-status-12-on-duplication-jobs/td-p/743854
I don't think you're in for any nasty surprises, and I reckon you're on the right track - you just have to set the value to be longer than the jobs were queued for (and reset it when you're done).
I also found this bit about that setting from the SLP Cheat Sheet for NetBackup 7.6 (https://www.veritas.com/support/en_US/article.000109501). The section below is from right near the bottom of that link:
An additional SLP tunable that is not included in the host properties but can be changed via bpsetconfig is
SLP.BIDFILE_CLEANUP_DELAY This defines how many days to keep a bidfile created by nbstserv. A bidfile contains the list of images to be processed for a duplication or replication job. The default is 3 days.
Couldn't find an equivalent doc for 7.7.x, but I don't know that they have updated the SLP Cheat Sheet.
Hope that helps,
Steve
10-12-2017 12:50 AM
Hi Rhino,
I just checked the applicable product list for the cheat sheet, and it says 7.6 and 7.7, so you should be good.
Steve
10-12-2017 01:41 AM
hi,
thanks for the links. seems this BIDFILE thing is a relatively unknown. the first link shows that no one is aware why BIDFILEs disappear after 3 days. a documentation failure on the part of NBU. how am i gonna explain this? i know someone will mention best practices but in the real world that doesn't always apply.
thanks again for confirming that the BIDFILE setting still works with 7.7.3
regards,
10-12-2017 02:52 AM
I also read through the 2015 post. At the time, not even Veritas Support engineer @mph999 knew about this timeout and cleanup.
Please bear in mind that only the duplication job has failed. The source backup is still intact as the retention level is set to infinity until all secondary tasks have completed.
SLP will automatically retry the failed duplication.
Hopefully @mph999 will be along soon to confirm.
10-12-2017 04:28 AM - edited 10-12-2017 04:29 AM
here's what's happening now. i did a manually backup of a failed scheduled backup last night. the SLP didn't kick in like i don't see any duplication job in activity monitor for this manual backup i did.
1. how to manual start SLP for the manual backup? yes there is an SLP that we've been using for years
2. so if only the duplication got cancelled, how do i apply another duplication if the backup images are still on disk?
edit: i restarted the NBU processes in the master and still no duplication job appearing for item 1 above.
regards,
Rino
10-12-2017 04:39 AM
You cannot manually start SLP-controlled duplication.
There are built-in rules for start and retry of duplications.
If job is considered to be a 'small' job (less than Minimum size per duplication job) then the default time to submit the duplication is 30 minutes.
Please have a look at your SLP parameters.
See https://www.veritas.com/support/en_US/article.v84182756_nbu_core_admin_web
10-12-2017 04:41 AM
this is a wild guess. the duplication jobs i'm seeing now are the sames ones that registerd an error 12 yesterday.
this was the error 12 list:
because the weekly SLP should only start on a weekend!
10-12-2017 04:46 AM
SLP Duplication schedule has nothing to do with the policy schedule.
From the screenshot, your SLP duplication window seems to be the default of 24 x 7.
Please check your SLP parameters as per my previous post.
10-12-2017 05:10 AM
my SLP parameters are the default. same as shown in that link. and as per my previous experience, duplication job for a finished backup should appear in the activity monitor after a few minutes.
i'll wait until tomorrow for the latest duplication job (for the manual backup done a few hours ago) to appear in the activity monitor.
10-14-2017 10:54 PM
10-23-2017 12:46 AM
Hopefully your 5 subsequent new threads answered your queries about duplications be resubmitted after a failure?
10-23-2017 01:05 AM
hopefully. still learning the ropes here. and i don't know, should i have sticked with a single thread?
04-27-2020 09:44 AM
we can run the manual duplication of image under SLP control in queued state by below commmand:
bpduplicate -lcm -dstunit <destination_storage_unit> -backupid <backup_id> -dp <volume_pool> -rl <retention>
-lcm defines its lifecycle managed