Checking cluster membership is not correct

[ivanov-aleksander]

I have faced and issue with a cluster with more than 10 nodes. I have created PR 30 to fix this issue.
But I found the second issue:
The condition that checks if node already joined the cluster is not correct.
For example:

1. Node mongooseim-10 fails to join to cluster because issue PR 30 as result pod will be restarted. But mnesia folder will be created.
2. Next start of the node mongooseim-10 skip joining the cluster because mnesia folder exist. As a result, it will be standalone node.

[michalwski]

Thanks for opening the issue. This is indeed a problem, I'll add this to our internal road map.