Reporting help to show staus codes for filesystem backups

[Melias]

Hi, We've recently migrated to Netbackup from EMC Networker and I need to recreate some audit reports in a similar format to the reports Networker used to create. Basically, I need to create a report that displays backup details for a historical date. The details I would like are the following: 1. Listing of filesystems that have been backed up and the status code indicating that all files on the particular filesystem have been successfully backed up and list any files that failed to be backed up. I only need to run/create the report for several clients. I'd like to make the report as simple as possible - it needs to be understood by auditors :).

The format of the existing Networker report is as follows:

Client Name Save Set Name Save Set ID Group Start Time Save Type Level Status

hostname /export/home 943503561 1/1/10 1:40:00 AM save 1 succeeded

hostname /global/.devices/node@1 926726350 1/1/10 1:40:00 AM save 1 succeeded

hostname /opt 641513783 1/1/10 1:40:00 AM save 1 succeeded

hostname /ora 960280774 1/1/10 1:40:00 AM save 1 succeeded

So I need to be able to prove to the auditos that the backup was successful and the required files successfully backed up for any date they require. I do not have any of the ops or reporting add-ons for Netbackup and this report is required before I could get the required approvals to buy/install any additional packages.

I'm hoping that since this is audit driven, someone else may have been down this path before.

Thank you.

[Andy_Welburn]

This may not give the depth of info you are looking for (e.g. filesystems) but could be a starting point.

e.g.

# bperror -d 10/09/2010 16:00:00 -e 10/10/2010 16:00:00 -t backstat -U
STATUS CLIENT        POLICY           SCHED      SERVER      TIME COMPLETED
  0         client1         Policy1         Schedule3   master       10/09/2010 23:46:47
196       client2         Policy2         Schedule3   master       10/10/2010 00:00:01
                 (client backup was not attempted because backup window closed)

[Ed_Wilts]

Depending on how you do the backup, this problem can get pretty hard with native tools, especially if you want historical data.

By default, NetBackup only keeps error summaries for 3 days so you can't find the failure status for jobs more than 3 days ago.  From an audit perspective, for say 2 weeks ago, you can't tell if the job was attempted and failed or if the job wasn't attempted at all.

If you basically want the list of successful files, you can certainly search the catalog or even more quickly, look at the STREAMS<policyname> files in the /usr/openv/netbackup/db/images/<client> directories.  Although not supported in any way by Symantec, the STREAMS files give you a good first look at what was successful (the definitive answer is in the catalog, NOT the STREAMS files).

The catalog gives you per-file information over time.  The STREAMS files give you per file-system information but only for the most recent backup.

For example, here's a quick look at a sample:

[root@osiris stp-aptare.bck.mrll.com]# pwd
/usr/openv/netbackup/db/images/stp-aptare.bck.mrll.com
[root@osiris stp-aptare.bck.mrll.com]# grep 'data01' STREAMSstp-aptare
4 1285981289 stp-aptare Monthly 0,0,2419200 0 0 /data01
4 1286609540 stp-aptare Weekly 0,0,604800 0 0 /data01
4 1286770347 stp-aptare Daily 1,0,86400 0 0 /data01

We then translate those Unix timestamps using your favorite utility.

[root@osiris stp-aptare.bck.mrll.com]# unixtime 1285981289 1286609540 1286770347
1285981289:     Fri Oct  1 20:01:29 2010
1286609540:     Sat Oct  9 02:32:20 2010
1286770347:     Sun Oct 10 23:12:27 2010
 

So I see that the last Daily was last night, the last weekly was Saturday, and the last Monthly was back on the 1st. 

Long-term, you'll find a commercial reporting tool VERY helpful for these kinds of questions - it will track the details about every job and store it in a database for you to run your reports on.  My preference is Aptare StorageConsole.  Just last week I was searching for details on tapes and jobs for retired systems - data that has long been flushed from the catalog - StorageConsole gave me exactly what I was looking for.

For auditing purposes, we typically demonstrate successful backups both by doing test restores on a regular basis and by doing catalog searches for specific files.  Not as good as StorageConsole, but it usually makes the auditors happy if you don't have anything better.

[RiaanBadenhorst]

I agree with Andy, you can keep a record of the daily jobs status using bperror. Should the auditors arrive and they want to see if a file is backed up, you can show them in the catalog, or you can generate a report of which files were backed up for which ever day they choose using the bplist command.

The information from bplist is always at hand, you dont need to generate it daily.

[Andy_Welburn]

DAILY basis as proof that the backup actually worked (via "Status of Backups" report) - they never asked for it again so I stopped killing forests!

[Ed_Wilts]

We have a script that scrapes the bperror output for failures, figures out what support team the error should go to (Unix admin, Windows admin, DBA, my team, etc.) and generates tickets via an email to our help desk who open up a ticket.  The individual teams then close the tickets when they've resolved the problems.

It's not perfect by a long stretch but it gives us an audit trail that the event was detected, logged, and resolved by a responsible party. 

We also have Aptare StorageConsole and we've never had an audit request we couldn't address.