On Sun, 2016-03-06 at 08:26 -0700, Shawn Heisey wrote: > On 3/5/2016 11:44 PM, YouPeng Yang wrote: > > We are using Solr Cloud 4.6 in our production for searching service > > since 2 years ago.And now it has 700GB in one cluster which is comprised > > of 3 machines with ssd. At beginning ,everything go well,while more and > > more business services interfered with our searching service .And a problem > > which we haunted with is just like a nightmare . That is the cpu sys > > usage is often growing up to over 10% even higher, and as a result the > > machine will hang down because system resources have be drained out.We have > > to restart the machine manually.
> One of the most common reasons for performance issues with Solr is not > having enough system memory to effectively cache the index. [...] How does this relate to YouPeng reporting that the CPU usage increases? This is not a snark. YouPeng mentions kernel issues. It might very well be that IO is the real problem, but that it manifests in a non-intuitive way. Before memory-mapping it was easy: Just look at IO-Wait. Now I am not so sure. Can high kernel load (Sy% in *nix top) indicate that the IO system is struggling, even if IO-Wait is low? YouPeng: If you are on a *nix-system, can you call 'top' on a machine and copy-paste the output somewhere we can see? - Toke Eskildsen, State and University Library, Denmark