Chris,

This sounds very suspiciously like a problem we had. We set up an 
experimental local ELK server (one node in the cluster) and fed it with 
logstash. I was manually cleaning up older data using the Elasticsearch 
Head plug-in, but over one weekend the cluster got into a funky state. The 
curl API said it was Yellow, but ES Head showed Green, and queries were 
hanging.

This was a VM that was dedicated to ES with 1TB disk space (only about 2% 
was ever used at any point in time), 4 CPUs, and 24GB RAM (though the Java 
JVM was not tuned to take advantage of all of this memory). Kibana was 
hosted as a site plug-in, but its usage was very light. Though I had been 
playing around with increasing the size limit of responses way past the 
default of 500, and I'm sure the ES server bore the brunt of that.

I stopped and restarted ES and everything went back to normal.

I installed Curator to clean up older indices automatically, and the 
problem has never returned. (I have also stopped telling Kibana to ask for 
up to 50000 response documents on a query!)

I suspect you're getting some sort of OOM condition and that's when things 
start looking odd.

Anyway, OOM is just a wild guess. I wouldn't have mentioned something so 
nebulous, but the symptoms you have are strikingly close to the ones we saw.

Brian

-- 
You received this message because you are subscribed to the Google Groups 
"elasticsearch" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to [email protected].
To view this discussion on the web visit 
https://groups.google.com/d/msgid/elasticsearch/ccf8ab3d-c89c-42da-95ba-1b25198fc445%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

Reply via email to