I recommend, in this case, that you use Solr's autocommit feature (see
solrconfig.xml) rather than having your indexing clients issue their
own commits. Overlapped searcher warming is just going to be too much
of a hit on RAM, and generally unnecessary with autocommit.
On Jul 30, 2009, at 2:28 PM, Jérôme Etévé wrote:
> Hi All,
> I'm planning to have a certain number of processes posting
> independently in a solr instance.
> This instance will solely act as a master instance. No clients
> queries on it.
> Is there a problem if i set maxWarmingSearchers to something like 30
> or 40?
> Also, how do I disable the cache warming? Is setting autowarmCount's
> to 0 enough?
> Jerome Eteve.
> Chat with me live at http://www.eteve.net >
> [hidden email]
: Is there a problem if i set maxWarmingSearchers to something like 30 or 40?
my personal opinion: anything higher then 3 indicates a serious
On a master, doing lots of updates, the "warming" time should be zero, so
there shouldn't ever be more then 2 searchers at one time -- 3 is being
generous incase you just happen to get some paralell rapid fire
add/commit pairs ... beyond that you're better off just letting any ohter
concurrent commit calls block for the few milli-seconds it will take to
finish the commit.
: Also, how do I disable the cache warming? Is setting autowarmCount's
: to 0 enough?
yes, but even better: make the cache sizes zero, that way if someone
accidently does query your master, you won't waste ram caching it.