On 3/17/2014 7:07 AM, adfel70 wrote: > we currently have arround 200gb in a server. > I'm aware of the RAM issue, but it somehow doesnt seems related. > I would expect search latency problems. not strange eofexceptions. > > regarding the http.timeout - I didn't change anything concerning this. > Do I need to explicitly set something different than the solr out-of-the-box > comes with? > > I'm also monitoring garbage collector metrics and I don't see anything > unsual..
Indexing puts extra strain on a Solr server, especially when that server does not have enough RAM to cache the entire index. The basic symptom is that *everything* takes longer than it normally does, including queries. A server that is indexing uses extra heap memory and does a LOT of I/O, both reading and writing. When you don't have enough RAM to cache the index effectively, the commit operation will compete with queries for space in the OS disk cache, making both operations slow. Your information says that GC is not a problem, but if it were a problem, indexing will make it many times worse. The specific EofException problem is because of the *client* ... not Solr. Whatever is talking to Solr is disconnecting before Solr is done, probably after 30 or 60 seconds. Solr and SolrJ do not configure timeouts by default, and neither does the Jetty server that is included with Solr. If you are using a load balancer, it's probably disconnecting there. When you mention "http.timeout" ... I actually have no idea what that is. I don't recall seeing a setting like that for Solr itself. It sounds like a setting for some other piece of software, perhaps a client, load balancer, or servlet container. Thanks, Shawn
