We had a weird incident happen and we stopped receiving log files for a very specific time window.
Is there a way to kick the UF clients to resend data to the indexers?
@michael_leo ,
You can resend old data by cleaning fishbucket index where Splunk software tracks how far into a file indexing has progressed, to enable the software to detect when data has been added to the file and resume indexing
Either you can remove the files under $SPLUNK_HOME/var/lib/splunk/fishbucket
for re-indexing all the inputs from that forwarder or you can use btprobe to selectively re-index specific files.
You need to stop the forwarder while doing the activities.
Have a look at this post for more information - how-to-reindex-data-from-a-forwarder
If you are re-indexing all the inputs, there could create duplicate events in your index and also this might count against your license.
FYI,
the clearing fishbucket trick does not seem to work on UFs... at least on 8.2
:/opt/splunkforwarder/etc/system/local# /opt/splunkforwarder/bin/splunk clean eventdata -index _thefishbucket
This action will permanently erase all events from the index '_thefishbucket'; it cannot be undone.
Are you sure you want to continue [y/n]? y
ERROR: Cleaning eventdata is not supported on this version.
Hi
another options are:
splunk btprobe -d SPLUNK_HOME/var/lib/splunk/fishbucket/splunk_private_db --file /var/log/access.log --resetYou should do this for all individual files which you want to reindexing
In both cases you must first stop UF before try above commands.
r. Ismo
Thanks!
I also discovered that the brute force technique works too:
rm -rf ../var/lib/splunk/fishbucket
Not as elegante, but effective...