Facing Error while upgrading elasticsearch cluster to 8.19.8

Elastic Stack Elasticsearch
1

Hey everyone,
I am currently upgrading my elasticsearch version from 8.11.3 to 8.19.8 and I am facing an error. I am using docker containers to deploy elasticsearch nodes. I was able to upgrade my coordinator and hot nodes but facing this error while upgrading my warm tier nodes.

Unable to create archive worker: unable to create native thread: possibly out of memory or process/resource limits reached | [0.003s][warning][os,thread] Failed to start thread "ArchiveWorkerThread" - pthread_create failed (EPERM) for attributes: stacksize: 1024k, guardsize: 4k, detached. | Error occurred during initialization of VM | Unable to create archive worker: unable to create native thread: possibly out of memory or process/resource limits reached exited with code 1

The warm nodes are on a machine which is using Centos release version 7.4.1708 and I have tried to set the nproc and ulimit to a higher value but to no avail. The node is not even starting up. But when I switched the node back to 8.11.3 it started and joined the cluster without any issues.

Is there any OS limitation with certain versions of Elasticsearch?

2

What does that mean? What did you specifically try? How did you conclude it was "to no avail"? That means nproc/ulimit were unable to be changed, or you just got the same error as before even with "new" values?

Note 8.19.8 uses JVM25 and 8.11.3 used JVM21 (assuming you are using bundled JVMs) - a fair bit will have changed.

3

I would recommend consulting the official support matrix.

1 Like