Dockerized Solr service stops responding

[Christian Spitzlay <ch...@biologis.com>]

Hi,

we are running Solr Cloud with a single instance of zookeeper and solr
in a docker setup.  We are experiencing occasional hangs where the solr 
service stops responding: clients time out but the docker containers are still
reported as healthy by the rancher administration UI.

"docker exec" can still connect to the solr container and open a shell but 
a local connect to solr using curl from inside the solr container times out, too.

A restart of the Solr container fixes the issue temporarily.

The last log entry we could find from before the hang is an exception
of the type I described in 
https://lists.apache.org/thread.html/eafcdc4c3c620f093f1292f9062badc4001b4957450ba69e4e668d1a@%3Csolr-user.lucene.apache.org%3E
(streaming expressions returning the correct result but throwing an exception anyway;
as described, we cannot switch to the export handler in all places as not all types we 
work with support docValues).

Can the two things be related, e.g. can lots of exceptions have a side effect
that leads to solr not responding any more?
Do you have any hints as to what to check for?


Best regards
Christian Spitzlay


--  

Christian Spitzlay
Senior Software-Entwickler

bio.logis Genetic Information Management GmbH
Zentrale:
Olof-Palme-Str. 15
D-60439 Frankfurt am Main

T: +49 69 348 739 116
christian.spitzlay@biologis.com
biologis.com

Geschäftsführung: Prof. Dr. med. Daniela Steinberger, Dipl.Betriebswirt Enrico Just
Firmensitz:  Altenhöferallee 3, 60438 Frankfurt am Main
Registergericht Frankfurt am Main, HRB 97945