You are viewing a plain text version of this content. The canonical link for it is here.
Posted to solr-user@lucene.apache.org by solr-user <so...@hotmail.com> on 2014/03/19 00:31:32 UTC

does shards.tolerant deal with this scenario?

hi all

I have some questions re shards.tolerant=true and timeAllowed=xxx

I have seen situations where shards.tolerant=true works; if one of the
shards specified in a query is dead, shards.tolerant seems to work and I get
results from the non-dead shards

However, if one of the shards goes down during the execution of a query, I
have to wait for the primary searcher (the solr sending the request to the
shards) to timeout, which can last minutes.  ie shards.tolerant doesn't seem
to work

question 1: is timeAllowed "shard-aware"?  ie in a sharded query, does this
param get used by all the shards specified or does it only get used by the
primary searcher?

question 2: Since shards.tolerant=true is not helping when a shard goes down
during query execution, is there any other way to deal with this?  If
timeAllowed is shard-aware, I would think that I could use timeAware and the
primary searcher would then wait xxx milliseconds and return with whatever
the other shards had sent back.  Is that correct?

thanks in advance




--
View this message in context: http://lucene.472066.n3.nabble.com/does-shards-tolerant-deal-with-this-scenario-tp4125300.html
Sent from the Solr - User mailing list archive at Nabble.com.