No master node reported

[Jorge_Mario_Montes_R (Jorge Mario Montes)]

This is on a cluster on cloud.elastic.co
After reducing memory form the master node, I'm unable to get it back running. Logs show:
master_not_discovered_exception
Tried to reapply the last successful configuration, but it doesn't work, keeps showing the same exception.

Another weird thing is that the new instance show 0/0 on storage

On this last image you can see the instance with 8g of RAM, but if I go to edit Configuration, it shows 16g:

Thanks in advance!

[warkolm (Mark Walkom)]

Welcome to our community!

Have you raise a request with the Support team about this? They should be able to look into what's happening and provide assistance.

[Jorge_Mario_Montes_R (Jorge Mario Montes)]

@warkolm

should be able to look into what's happening and provide assistance.

Hi, yes, but no answer yet. This is a production env so posted in here to see if anyone knows about it.

[warkolm (Mark Walkom)]

Fair enough. We don't have access to the backend stuff that the Support team does, so it's hard for us to troubleshoot given what you are/aren't seeing sorry to say

[Jorge_Mario_Montes_R (Jorge Mario Montes)]

Also, I can't access snapshots. Is there a way to force recreating the master instance?
The data instance looks fine.
Or is there a way I can create a new deployment using the data instance?

[Jorge_Mario_Montes_R (Jorge Mario Montes)]

Am I missing something? Is ECE meant for non-production environments? Been down for 72 hours and still no response about the ticket I sent to support.
Is there some kind of additional fee for support? I don't see any upgrade option on Elastic Coud UI.

[Alex_Piggott]

If you click on the three dots in the top right of instance you can force override its instance size back to 16GB (assuming there is capacity on the allocator - if not you may need to clear some), and then once it's back to life the plan should succeed

Can you DM me the support ticket id, I'll have a look at why you didn't hear anything back

Alex

(Incidentally needing >8GB on your master nodes is not that common and can be the sign that you are either oversharded or have a "mapping leak", it's worth looking into once you stabilize the cluster)