Hello,
is it possible to set up a Universal Forwarder in such a way that it uses one indexer, and will try to send it's data to a second configured indexer only when the first one is no longer available?
I generally need to only use one indexer. I do not want load balancing for now.
And I do not need/want to clone my data, for license reasons.
Thanks,
Christoph
To the best of my knowledge there is no out-of-the-box solution for this.
Perhaps you can make it work through some DNS trickery or have local script modifying the hosts
file of the operating system. Maybe.
If you have the extra hardware standing by, why not use both?
UPDATE:
Well, I actually realized that it might be as simple as setting the AutoLBFrequency
in outputs.conf on the forwarder to a VERY high value (billions). Then the forwarder should not switch to the alternate indexer unless the primary goes down.
Unfortunately, the forwarder would not switch back automatically when the primary is available again. You'd have to manually restart the alternate indexer, thereby terminating the sessions, which would cause the forwarders to return to the primary indexer.
This is a pretty weird approach, since you still need to have the alternate indexer up-and-running (or at least in hot standby) at all times.
/k
this is officially supported in splunk 6.6.0.
To the best of my knowledge there is no out-of-the-box solution for this.
Perhaps you can make it work through some DNS trickery or have local script modifying the hosts
file of the operating system. Maybe.
If you have the extra hardware standing by, why not use both?
UPDATE:
Well, I actually realized that it might be as simple as setting the AutoLBFrequency
in outputs.conf on the forwarder to a VERY high value (billions). Then the forwarder should not switch to the alternate indexer unless the primary goes down.
Unfortunately, the forwarder would not switch back automatically when the primary is available again. You'd have to manually restart the alternate indexer, thereby terminating the sessions, which would cause the forwarders to return to the primary indexer.
This is a pretty weird approach, since you still need to have the alternate indexer up-and-running (or at least in hot standby) at all times.
/k
Glad it helped, even though I'm a bit curious to the reasons for the requirement.
Thank you! This is indeed weird but interesting, and might actually work for me. Thank you for sharing your insight!
:-)
see update above