Open uriariel opened 4 years ago
Hi @uriariel ,
The response seems to indicate the ElasticSearch component is working in read-only mode - this might be caused by several reasons, among them are resource issues (i.e. insufficient storage on disk) or a migration issue.
Two questions:
docker logs trains-elastic >& elastic_logs.txt
I installed the server and used it with no issues until shutting it down after a few hours. (with docker-compoe down) When I tried to re run it I ranned into this issue.
I'm sorry, I solved the problem by removing all of the data directories and recreating them, so I don't think that the current logs will be of any help.
Just a question - is it possible the previous server you installed (and than took down) was v0.15.1 (or earlier) and the newly installed server is v0.16.0?
I don't think so, I didn't touch the docker-compoe.yml file.
OK, thanks for your time 👍 You're welcome to close this issue if the problem is currently solved (you can always reopen if it happens again...)
After another tackle with this issue, I understood that the problem occurs when the host running the docker is dealing with almost full storage. Elasticsearch has a feature that switches it into ease only mode if the server has storage issues.
I just cleaned up a bit and it got back to normal.
I see. See my last comment here for more info on how to configure ES (via our docker-compose) for different watermarks. However, the best option is more disk space 😄
Thanks
Trains server running on docker compose isn't able to perform experiments and prompts the next error:
this is the errors from trains integration in my code:
And this is the errors I get from docker-compose logs: