-
Notifications
You must be signed in to change notification settings - Fork 174
Nodes page not showing JVM usage for one of the nodes #1673
Comments
After posting this I got an error stacktrace posted below:
|
@evanccnyc Do you get that every time? I am unable to reproduce this with two Graylog server nodes. |
Not everytime. I had to do a reinstall for unrelated reasons and it seems to have gone away. |
Thank you for the feedback! I am closing this now. Please re-open or create a new ticket if you run into this again. Thank you! |
This has reappeared. |
Log from web interface
|
In one of the server nodes logs I saw this:
Once i ran ntpdate I was able to access the node web interface again. |
Can we reopen this bug? |
I will re-open the issue. Thanks for the updates. |
I am also getting this issue, alongside a similar one to #1517 where randomly if I try access the nodes page, I get "Oops, an error occured". It only appeared today, after the cluster has been running fine for over 24 hours. I have 2 nodes, node 1 (master) which works fine and the jvm stats are correct, then the second node which displays JVM stats sometimes, and other times it's blank. If I turn off the second node (The one not displaying the stats) then it works fine, even going through the second node's web address to view it. In the Web server logs when I receive the error I am seeing;
There doesn't appear to be any errors when the JVM stats don't work. I am also seeing the My boxes are running with Graylog 1.1.6, Elasticsearch 1.3.9, CentOS7, Java 1.8 |
@MattDevUK The NodePingThread warning means one of the nodes is having garbage collection (or other) pauses of over 5 seconds, which leads to all kinds of different errors. |
@kroepke I've tripled the heapsize and checked with jmap, it's using about 50% of the space allocated but still getting that warning in the logs every second. also, only the second node is getting this and it's the exact same setup as the master. |
Graylog: 1.2.2
On a brand new setup on the nodes page one of the nodes doesnt show JVM usage like the other one does.
If I click on the actual node, I see the data however.
Screenshot attached:
The text was updated successfully, but these errors were encountered: