lucene-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Upayavira">
Subject Re: Load balancing within distributed search
Date Tue, 01 Mar 2011 14:14:54 GMT

On Tue, 01 Mar 2011 13:18 +0000, "Alex Cowell"
<> wrote:

  Hi Upayavira,
  I believe there is a basic implementation of load balancing
  for distributed search. The submit() method in the
  HttpCommComponent class within SearchHandler contains code
  which splits a pipe character delimited string of shards and
  uses a LoadBalancedHttpSolrServer to send requests to them.
  But the functionality is limited to the features of the
  LBHttpSolrServer, such as round-robin choosing of shards to
  query (although when the string of |-delimited shards is
  split, the resulting individual shards are shuffled), so I'm
  sure this could be improved upon.

Yes, thanks Alex, I see it now. My checkout of trunk was old, and
clearly predated that code.

I guess my question then is would a patch be accepted to backport
it to the 3.x branch?



On Tue, Mar 1, 2011 at 11:20 AM, Upayavira <[1]>

  The SolrCloud doc suggests a request a shard syntax like this:
  Where a pipe character (|) is used to provide load balancing
  for shards. I've looked in trunk but cannot find anything that
  like it.
  Has this been implemented anywhere? If not, I'd like to have a
  go at
  implementing it.
  The requirements as I see it are:
   * randomised choice
   * on an exception, try other hosts for that shard, until
  More complex would be:
   * if a host is down, note it as such and try it again only
  after a
   pre-defined interval.
  Any thoughts?
  Enterprise Search Consultant at Sourcesense UK,
  Making Sense of Open Source
  To unsubscribe, e-mail: [2]
  For additional commands, e-mail: [3]


Enterprise Search Consultant at Sourcesense UK, 
Making Sense of Open Source

View raw message