# Elasticsearch unstable cluster **URL:** \ **Category:** Elasticsearch\ **Created:** [December 31, 2023, 1:34pm UTC](https://discuss.elastic.co/t/elasticsearch-unstable-cluster/350157 "2023-12-31T13:34:30Z")\ **Posts on this page:** 8\ **Page:** 2
**Author:** ![elasticexpert](https://avatars.discourse-cdn.com/v4/letter/e/a5b964/32.png) [@elasticexpert](https://discuss.elastic.co/u/elasticexpert)\ **Post date:** [January 21, 2024, 2:08pm UTC](https://discuss.elastic.co/t/elasticsearch-unstable-cluster/350157/28 "2024-01-21T14:08:09Z")
No, I don't have any custom non-standard config. And yes the nodes are running directly on the host ---
**Author:** ![Christian\_Dahlqvist](https://sea2.discourse-cdn.com/elastic/user_avatar/discuss.elastic.co/christian_dahlqvist/32/4617_2.png) [@Christian\_Dahlqvist](https://discuss.elastic.co/u/Christian_Dahlqvist)\ **Post date:** [January 21, 2024, 2:23pm UTC](https://discuss.elastic.co/t/elasticsearch-unstable-cluster/350157/29 "2024-01-21T14:23:45Z")
Elasticsearch by default assumes it is the only process running on the host (best practice) and sizes threadpools based on this. This probably means that your threadpools are 8 times as large as they should be, which may be causing issues. As recommended [in the documentation](https://www.elastic.co/guide/en/elasticsearch/reference/8.12/modules-threadpool.html#node.processors) (which describes your scenario) you should set the `node.processors` for all your node, probably to `8` or `9`. As noted in the docs, this is considered an expert setting and other settings may need to also be adjusted. I am not familiar with this so will leave that for someones else to guide on. ---
**Author:** ![DavidTurner](https://sea2.discourse-cdn.com/elastic/user_avatar/discuss.elastic.co/davidturner/32/22453_2.png) [@DavidTurner](https://discuss.elastic.co/u/DavidTurner)\ **Post date:** [January 21, 2024, 3:25pm UTC](https://discuss.elastic.co/t/elasticsearch-unstable-cluster/350157/30 "2024-01-21T15:25:49Z")
> [@elasticexpert](#): > > the nodes are running directly on the host Yeah as well as the sizing problems that Christian mentioned I wouldn't be surprised to see all sorts of weird contention effects if you just simply run lots of ES nodes on a big NUMA system like this. Tracking down and fixing stuff like this is more of a sysadmin/Cisco setup question than anything specific to ES so you may need to look for a different forum if you need help with that. But basically if something as simple as `top -n1` takes tens of seconds to run then there's something deeply wrong with your system. ---
**Author:** ![elasticexpert](https://avatars.discourse-cdn.com/v4/letter/e/a5b964/32.png) [@elasticexpert](https://discuss.elastic.co/u/elasticexpert)\ **Post date:** [January 24, 2024, 7:56am UTC](https://discuss.elastic.co/t/elasticsearch-unstable-cluster/350157/33 "2024-01-24T07:56:26Z")
What do you think would be the best practice to install elasticsearch cluster on my servers? Using Docker maybe? ---
**Author:** ![DavidTurner](https://sea2.discourse-cdn.com/elastic/user_avatar/discuss.elastic.co/davidturner/32/22453_2.png) [@DavidTurner](https://discuss.elastic.co/u/DavidTurner)\ **Post date:** [January 24, 2024, 8:09am UTC](https://discuss.elastic.co/t/elasticsearch-unstable-cluster/350157/34 "2024-01-24T08:09:34Z")
![image](https://us1.discourse-cdn.com/elastic/original/3X/3/d/3d808011ac35f847a760728a2dacd3374abf2f91.png) I'm not sure exactly why you keep on deleting your posts and and then recreating them a few hours/days later, but I suspect you might be doing this to send a duplicate notification to the folks in this thread in the hope of getting more of their attention. If so, that's rather rude, and could be considered a breach of the [community code of conduct](https://www.elastic.co/community/codeofconduct). We're all just volunteers here, and this kind of behaviour makes us less likely to choose to spend our limited time with you on this question and other questions you might have in future. I don't have anything more to add here unfortunately, you need to talk to someone with experience of system administration for the system you're using. Your problems aren't really anything to do with Elasticsearch itself, so while it's possible that there's anyone with that experience in this forum, it's also fairly unlikely. ---
**Author:** ![elasticexpert](https://avatars.discourse-cdn.com/v4/letter/e/a5b964/32.png) [@elasticexpert](https://discuss.elastic.co/u/elasticexpert)\ **Post date:** [January 24, 2024, 8:29am UTC](https://discuss.elastic.co/t/elasticsearch-unstable-cluster/350157/35 "2024-01-24T08:29:36Z")
No its not what i meant to do, its just because when i do reply to someone in the first time its not showing this: ![image](https://us1.discourse-cdn.com/elastic/original/3X/6/8/68b0a3e4cf2e28bd53a27ab7ae0d5b317cc20558.png) Sorry if it looked weird 🙏 EDIT: just to show you (I did a reply on your comment and its not showing): ![image](https://us1.discourse-cdn.com/elastic/original/3X/c/c/cc01ec8cde2b60a9c5a91a003871065c2a9566ee.png) ---
**Author:** ![DavidTurner](https://sea2.discourse-cdn.com/elastic/user_avatar/discuss.elastic.co/davidturner/32/22453_2.png) [@DavidTurner](https://discuss.elastic.co/u/DavidTurner)\ **Post date:** [January 24, 2024, 9:55am UTC](https://discuss.elastic.co/t/elasticsearch-unstable-cluster/350157/36 "2024-01-24T09:55:18Z")
Ok, that's expected, you only get this indication if you're replying to a post before the last one in the thread. Please stop doing this in any case. ---
**Author:** ![system](https://us1.discourse-cdn.com/elastic/original/3X/1/a/1ac57faf039f6b580b3f104ef42a2a89e41014de.png) [@system](https://discuss.elastic.co/u/system)\ **Post date:** [February 21, 2024, 9:56am UTC](https://discuss.elastic.co/t/elasticsearch-unstable-cluster/350157/37 "2024-02-21T09:56:01Z")
Closed on February 21, 2024, 9:56am UTC This topic was automatically closed 28 days after the last reply. New replies are no longer allowed. [Previous page](https://discuss.elastic.co/t/elasticsearch-unstable-cluster/350157.md?page=1)