Getting Data In

training splunk to recognize events from bacula

Communicator

I am running into trouble getting splunk to properly break down events from bacula. Below is an example of a bacula event. The below represents a single backup job in the bacula log. I thought it would be easy to setup since every backup job has a common JobId. I thought I would be able to give splunk a line_breaker and then a regex but that is not working. Do I need to use the transaction command instead? I appreciate any help since I am getting no where fast.

Thanks

Ed

14-Mar 04:09 xxx-dir JobId 211: Start Backup JobId 211, Job=xxx.2011-03-14_04.05.00_29
14-Mar 04:09 xxx-dir JobId 211: Using Device "FileStorage"
14-Mar 04:09 xxx-sd JobId 211: Volume "Vol0007" previously written, moving to end of data.
14-Mar 04:09 xxx-sd JobId 211: Ready to append to end of Volume "Vol0007" size=36444353196
14-Mar 04:09 xxx-sd JobId 211: Job write elapsed time = 00:00:01, Transfer rate = 1.871 M Bytes/second
14-Mar 04:09 xxx-dir JobId 211: Bacula xxx-dir 5.0.3 (30Aug10): 14-Mar-2011 04:09:16
  Build OS:               i686-redhat-linux-gnu redhat Enterprise release
  JobId:                  211
  Job:                    chi01fep110.2011-03-14_04.05.00_29
  Backup Level:           Incremental, since=2011-03-13 04:07:25
  Client:                 "xxx-fd" 5.0.3 (30Aug10) i686-redhat-linux-gnu,redhat,Enterprise release
  FileSet:                "standard_etc" 2011-03-03 23:05:00
  Pool:                   "File" (From Job resource)
  Catalog:                "MyCatalog" (From Client resource)
  Storage:                "File" (From Job resource)
  Scheduled time:         14-Mar-2011 04:05:00
  Start time:             14-Mar-2011 04:09:16
  End time:               14-Mar-2011 04:09:16
  Elapsed time:           0 secs
  Priority:               10
  FD Files Written:       784
  SD Files Written:       784
  FD Bytes Written:       1,788,470 (1.788 MB)
  SD Bytes Written:       1,871,767 (1.871 MB)
  Rate:                   0.0 KB/s
  Software Compression:   64.7 %
  VSS:                    no
  Encryption:             no
  Accurate:               no
  Volume name(s):         Vol0007
  Volume Session Id:      146
  Volume Session Time:    1299264299
  Last Volume Bytes:      36,446,249,241 (36.44 GB)
  Non-fatal FD errors:    0
  SD Errors:              0
  FD termination status:  OK
  SD termination status:  OK
  Termination:            Backup OK

14-Mar 04:09 xxx-dir JobId 211: Begin pruning Jobs older than 1 month .
14-Mar 04:09 xxx-dir JobId 211: No Jobs found to prune.
14-Mar 04:09 xxx-dir JobId 211: Begin pruning Jobs.
14-Mar 04:09 xxx-dir JobId 211: No Files found to prune.
14-Mar 04:09 xxx-dir JobId 211: End auto prune.
Tags (1)
0 Karma

SplunkTrust
SplunkTrust

Hi Ed,

In this instance I personally would probably treat this as several events and use transaction as needed to weld them back together. I would probably use a BREAK_ONLY_BEFORE in props.conf such that lines with a date/time on them delinate events.

[bacula]
BREAK_ONLY_BEFORE=^\d{2}-[A-Za-z]{3}\s+\d{2}:\d{2}\s+
SHOULD_LINEMERGE = true
TIME_FORMAT=%d-%b %H:%M
TIME_PREFIX=^
MAX_TIMESTAMP_LOOKAHEAD=13

Put that in your props.conf for this sourcetype. If it works right, your multi-line event should be treated as a single event. This will only take affect on data indexed after the change is made.

UPDATE - Ed, please try the above props.conf entry - it worked properly for me on your test data. Timestamps came out correct, and lines were delineated/merged appropriately.

SplunkTrust
SplunkTrust

Ed, see update above....

0 Karma

Communicator

The event is still being broken up into pieces not related to the date/time. I added new data to the log after adding the above to the props.conf. Any ideas? Thanks!

0 Karma

Communicator

Will give this a try and get back to you - Thanks!

0 Karma