Netbackup Detailed Status

[Sid1987]

Hi Guys,

 I am in process of writing a script for checking a particular policy, if the jobs ran successfully or not. I will be comparing with number of files backed up previous date. I will do that. However I am also looking if the job completed partially, I would want to get the list of files skipped or gave error, which is present in detailed status of the job in Activity Monitor. I can get that in command line through bpdbjobs -all_columns. But my problem is how do I organise the output of this command to get only the lines which gave error or were skipped, as this command's output is single line I guess.

Thanks

Sid

[RamNagalla]

detail status does not all the files if too many files are skipped...

I would suggest you to use the opscenter report better results...

Opscenter has ablity to list out the skipped files for each job also the can automate the reports that you are trying to configure..

[RiaanBadenhorst]

Hi Sid,

Take a look at https://support.symantec.com/en_US/article.TECH65123.html it gives all the columns of bpdbjobs -most_colums.

You could grab the following columns and report based on that.  Assuming you're running unix/linix and can script :)

1 jobid
2 jobtype
3 state
4 status
5 class
6 schedule
7 client
8 server
9 started
10 elapsed
11 ended
12 stunit
13 try
14 operation
15 kbytes
16 files
17 pathlastwritten
18 percent
19 jobpid
20 owner
21 subtype
22 classtype
23 schedule_type
24 priority
25 group
26 masterserver
27 retentionunits
28 retentionperiod
29 compression
30 kbyteslastwritten
31 fileslastwritten
32 trystatus
33 trystatusdescription
34 parentjob
35 kbpersec
36 copy
37 robot
38 vault
39 profile
40 session
41 ejecttapes
42 srcstunit
43 srcserver
44 srcmedia
45 dstmedia
46 stream
47 suspendable
48 resumable
49 restartable
50 datamovement
51 snapshot
52 backupid
53 killable
54 controllinghost

You could also just pull this out of opscenter.

[sdo]

Sid, over the years there have been various examples posted in different places on the internet of using perl to process an '-all_columns' file.  Yes, it is tricky, due to the two embedded tables within each record.  And, as Ram has noted, not all skipped/missed files are logged anyway (in bpdbjobs).

Yes, all 'activity' monitor data is all on one record/line per job - thus records can get very large indeed - sometimes several MB in length, with thousands of fields.  Also watch out for embedded (escaped) commas within the text, and sometimes trailing empty lines are also escaped - and you'll also have to handle commas embedded within quoted strings.

And bperror cannot be used, as this notes the suppressed messages too.

Clients do have an "Error_Messages_Max" setting, which is typically defaulted to 10.  So, you could try increasing this to a larger number to receive more messages re missed files - but I'd be unsure of any recommended maximum for this setting.

I'm not 100% convinced that OpsCenter will have the full list of all skipped files, because I think it is a client side setting which limits the list size, i.e. limits the number of messages sent back to the master server - and I'm not aware of OpsCenter polling clients to retrieve the full list of all missed files.  So, maybe the full list is only ever in the client side bpbkar or client side bpbrm log files.

[RiaanBadenhorst]

Hi Sid

Have you checked the contents of the domain_SkippedFileArchive in OpsCenter. According to the OpsCenter Schema doc "This table contains the information of files being skipped in the backup.". It has a Column Name called "Name" >>> Specifies name of skipped file

I've tried to select everything from this table on my installation but its blank. Must be because my backups work perfectly, or maybe because I have 1 client and I backup /etc. LOL

See if you get some information from it via opscenter. Let me know.

[sdo]

Searching the OpsCenter forum for "skipped files", produces several hits:

https://www-secure.symantec.com/connect/search/apachesolr_search/skipped%20files?filters=type%3Asc_forum%20tid%3A711%20tid%3A1601&retain-filters=1

[Sid1987]

Hi All,

Thanks for the response, I agree with sdo. Moreover I don't want use Opscenter as I don't want emails on daily basis. My aim is to check a particular policy if firstly backup has completed successfully by comparing it number of files backed up, as it is a client driven backup and filesystems get mounted on my media server and backup is triggered through my script. Now many a times all the filesystems are not mounted but still backup completes with status 0 but it is not a complete backup. So I need to compare each days backp with a certain no of files(I am still working on finding a profound way). So 1st task is to check if backup completed with status 0 and it backed up approximately correct size of data. In case not then i want to check error and trigger an email

Now i want skipped files or erroneous files if backup partially completes. I don't want all the files just few files for respective team to check

@sdo, can you point me to the perl you are talking in your comment

Thanks

Sid

[sdo · Accepted solution]

Does this help?

http://www.backupcentral.com/phpBB2/two-way-mirrors-of-external-mailing-lists-3/symantec-netbackup-18/bpdbjobs-all-columns-59816/

.

And look at Bob Stumpr's description of the structure of the embedded lists here - not the use of [] to denot structures:

http://www.tek-tips.com/viewthread.cfm?qid=855463

...and it looks like I'd forgotten that there are actually three embedded tables.

.

If you want to check whether a file system was backed-up - I would probably use 'bplist' with some judicious use of CLI switches - to detect whether a top-level (or MFD) folder was present in a lisiting for a given time period - or translate a backup ctime to a specific date-time and check specific images by specifying the same begin and end time on the bplist CLI.

[Sid1987]

Thanks sdo

http://www.backupcentral.com/phpBB2/two-way-mirrors-of-external-mailing-lists-3/symantec-netbackup-18/bpdbjobs-all-columns-59816/

this did it. attaching the perl script i edited it for a particular jobid and it works.

Sid

[sdo]

Thanks for sharing the code.