Souad90
Enthusiast
Enthusiast

Cluster going offline takes too much time

Jump to solution

Hi everyone,

I turn off the cluster in order to restart vRops but it's taking way too much time. The reason I want to restart is because the UI is so slow. Does any one had this issue before?

What would be the problem ?

---EDIT ---

After waiting for a couple of hours, the cluster shows :Cluster failed to go offline

Capture.PNG

Thank you very much.

Regards.

1 Solution

Accepted Solutions
sxnxr
Commander
Commander

As far as i am aware (and i am willing to be corrected) It depends how many nodes you have, the number of metrics, how utilised each node is and so on will contribute to how long the length of time to shut down the cluster is. Vrops databases are in memory (6.1 uses Cassandra 6.0 uses gemfire). when shutting down the cluster the in memory databases will be saved to disk or flushed or what ever it is called so you can safely shut down the nodes. I had time outs and so on when shutting down nodes because i didn't have them sized correctly. Stopping the adapters from collecting could help take the cluster offline faster.

In my environment i have 4 nodes and 2 remote collectors collecting almost 2m metric and it can take up to 45 mins to take the cluster offline.

View solution in original post

4 Replies
virtualg_uk
Leadership
Leadership

Is it possible to take one node offline and restart the others to isolate the issue?

__________________________________________________________ Graham User Moderator http://virtualg.uk | @VirtualG_UK
0 Kudos
Souad90
Enthusiast
Enthusiast

We actually always restart all the cluster. we have 4 nodes. and after clicking on Take offline again, it turned offline quickly. Taking online did not take much time as well after. But I still want to know why it was like that.

0 Kudos
sxnxr
Commander
Commander

As far as i am aware (and i am willing to be corrected) It depends how many nodes you have, the number of metrics, how utilised each node is and so on will contribute to how long the length of time to shut down the cluster is. Vrops databases are in memory (6.1 uses Cassandra 6.0 uses gemfire). when shutting down the cluster the in memory databases will be saved to disk or flushed or what ever it is called so you can safely shut down the nodes. I had time outs and so on when shutting down nodes because i didn't have them sized correctly. Stopping the adapters from collecting could help take the cluster offline faster.

In my environment i have 4 nodes and 2 remote collectors collecting almost 2m metric and it can take up to 45 mins to take the cluster offline.

View solution in original post

ahmad090
Enthusiast
Enthusiast

HI

i have the same issue , but i have only 1 node , with same problem even after rebooting the problem persist s.png

0 Kudos