[
https://issues.apache.org/jira/browse/SOLR-9824?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15733911#comment-15733911
]
Mark Miller commented on SOLR-9824:
-----------------------------------
So what I do is use a max int poll time and take advantage of that fact that we
know internally we use CUSS like:
* Create CUSS
* AddDocs
* BlockUntilFinished
* Done
This is not the general case where we may then add more docs or something, when
we call blockUntilFinished we know all the docs we are going to add are in the
queue.
So we use a max int queue poll time - we want a single connection and we know
how long we want to keep the connection up. That means we have to bail on those
queue waits though - so in blockUntilFinished, we just wait until the queue has
been emptied and then we interrupt the waits on the queue polling so that all
the runners shutdown.
The idea is, rather than this lastDoc flag that has so many issues, we just try
to keep polling through the length of the request.
> Documents indexed in bulk are replicated using too many HTTP requests
> ---------------------------------------------------------------------
>
> Key: SOLR-9824
> URL: https://issues.apache.org/jira/browse/SOLR-9824
> Project: Solr
> Issue Type: Improvement
> Security Level: Public(Default Security Level. Issues are Public)
> Components: SolrCloud
> Affects Versions: 6.3
> Reporter: David Smiley
> Attachments: SOLR-9824.patch
>
>
> This takes awhile to explain; bear with me. While working on bulk indexing
> small documents, I looked at the logs of my SolrCloud nodes. I noticed that
> shards would see an /update log message every ~6ms which is *way* too much.
> These are requests from one shard (that isn't a leader/replica for these docs
> but the recipient from my client) to the target shard leader (no additional
> replicas). One might ask why I'm not sending docs to the right shard in the
> first place; I have a reason but it's besides the point -- there's a real
> Solr perf problem here and this probably applies equally to
> replicationFactor>1 situations too. I could turn off the logs but that would
> hide useful stuff, and it's disconcerting to me that so many short-lived HTTP
> requests are happening, somehow at the bequest of DistributedUpdateProcessor.
> After lots of analysis and debugging and hair pulling, I finally figured it
> out.
> In SOLR-7333 ([~tpot]) introduced an optimization called
> {{UpdateRequest.isLastDocInBatch()}} in which ConcurrentUpdateSolrClient will
> poll with a '0' timeout to the internal queue, so that it can close the
> connection without it hanging around any longer than needed. This part makes
> sense to me. Currently the only spot that has the smarts to set this flag is
> {{JavaBinUpdateRequestCodec.unmarshal.readOuterMostDocIterator()}} at the
> last document. So if a shard received docs in a javabin stream (but not
> other formats) one would expect the _last_ document to have this flag.
> There's even a test. Docs without this flag get the default poll time; for
> javabin it's 25ms. Okay.
> I _suspect_ that if someone used CloudSolrClient or HttpSolrClient to send
> javabin data in a batch, the intended efficiencies of SOLR-7333 would apply.
> I didn't try. In my case, I'm using ConcurrentUpdateSolrClient (and BTW
> DistributedUpdateProcessor uses CUSC too). CUSC uses the RequestWriter
> (defaulting to javabin) to send each document separately without any leading
> marker or trailing marker. For the XML format by comparison, there is a
> leading and trailing marker (<stream> ... </stream>). Since there's no outer
> container for the javabin unmarshalling to detect the last document, it marks
> _every_ document as {{req.lastDocInBatch()}}! Ouch!
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscr...@lucene.apache.org
For additional commands, e-mail: dev-h...@lucene.apache.org
