Wait for Elasticsearch to start should be modified

[afeefghannam89]

TASK [netways.elasticstack.elasticsearch : Fail if cluster is not ready yet] ****************************************************************************************************************************************************************************************************************
fatal: [elasticcollection-1]: FAILED! => {"changed": false, "msg": "Elasticsearch cluster is not ready (status: yellow , yet. Please rerun again later."}
fatal: [elasticcollection-2]: FAILED! => {"changed": false, "msg": "Elasticsearch cluster is not ready (status: yellow , yet. Please rerun again later."}
fatal: [elasticcollection-3]: FAILED! => {"changed": false, "msg": "Elasticsearch cluster is not ready (status: yellow , yet. Please rerun again later."}

[widhalmt]

What do you suggest? Longer timeouts?

[afeefghannam89]

I suggest to use curl or use until as a parameter with more tries for example 5

[afeefghannam89]

It is the same issue here #100 When I rerun the playbook then the task is ok. The message in log say it is authentication failure, what make man wondering. What do you think about resolving first the timeout issue, then see if the authentication problem exists?

[widhalmt]

Yes, sound like a good idea! Please don't use curl with command or shell. If we want to check directly, maybe get_url or similar modules work. If they don't, maybe it's time to start to build a module to check the status of Elasticsearch?

[widhalmt]

Well... considering the code we already have, maybe using curl won't be as bad. ;-) I'm ok with staying with curl for now and planning a better solution for all curls we have for a future rework.

[widhalmt]

Thanks for the suggestion with until. That makes the whole code a lot better, I guess.