How use disk-aware allocator setting?

Elastic Stack Elasticsearch
1

Howdy! I'd love to use the recently added support for disk usage in the desired balance allocator as I just upgraded to 8.6 in Elastic Cloud.

However, it looks like by default cluster.routing.allocation.balance.disk_usage is set to a neutral setting. I was wondering if there was recommendation for raising this and what value I should set it to (in yaml)? Due to differing levels of traffic to different indices, our nodes are very disk-space unbalanced and I'd like to raise it's priority when rebalancing.

Edit: It looks like elastic cloud does not allow me to touch this setting, so this question may be moot.

2

The default is supposed to be good, but this feature is new so may need refinement, especially in more unusual cases. Could you send me your deployment ID and we'll see if we can learn anything from your cluster? A DM is fine if you don't want to share publicly.

1 Like
3

I manually reallocated some larger shards and things are more balanced again, it may have just been that the allocator didn't see reason to shift things around.

I'll keep an eye on it moving forward to see if it gets uneven again, and DM-ing deployment ID in case there's any interesting data there.

Edit: I think you may have forum DM's turned off :sweat_smile:

4

Hm I didn't think so, at least I've not changed anything and don't see anything in my preferences about that. @warkolm could you check for me?

5

Try now @RichiCoder.

6

This topic was automatically closed 28 days after the last reply. New replies are no longer allowed.