We do. Indexing is always running. Fix version is 6.3 so can I assume that the issue is fixed in 6.3? We are running 6.3 right now so or the fix is not in 6.3 or another issue is causing the full resync.

BR
Arkadi


On 25-11-16 18:23, Pushkar Raste wrote:

Did you index any documents while node was being restarted? There was a issue introduced due to IndexFingerprint comparison. Check SOLR-9310. I am not sure if fix made it to Solr6.2


On Nov 25, 2016 3:51 AM, "Arkadi Colson" <ark...@smartbit.be <mailto:ark...@smartbit.be>> wrote:

    I am using SolrCloud on version 6.2.1. I will upgrade to 6.3.0
    next week.

    This is the current config for numVersionBuckets:

        <updateLog>
          <str name="dir">${solr.ulog.dir:}</str>
          <int
    name="numVersionBuckets">${solr.ulog.numVersionBuckets:65536}</int>
        </updateLog>

    Are you saying that I should not use the config below on SolrCloud?

      <requestHandler name="/replication" class="solr.ReplicationHandler">
        <lst name="defaults">
          <str name="maxWriteMBPerSec">18.75</str>
          <str name="commitReserveDuration">05:00:00</str>
          <str name="httpConnTimeout">150000</str>
          <str name="httpReadTimeout">300000</str>
        </lst>
      </requestHandler>

    Br,
    Arkadi


    On 24-11-16 17:46, Erick Erickson wrote:

        Hold on. Are you using SolrCloud or not? There is a lot of
        talk here
        about masters and slaves, then you say "I always add slaves
        with the
        collection API", collections are a SolrCloud construct.

        It sounds like you're mixing the two. You should _not_ configure
        master/slave replication parameters with SolrCloud. Take a
        look at the
        sample configs....

        And you haven't told us what version of Solr you're using, we can
        infer a relatively recent one because of the high number you
        have for
        numVersionBuckets, but that's guessing.

        If you are _not_ in SolrCloud, then maybe:
        https://issues.apache.org/jira/browse/SOLR-9036
        <https://issues.apache.org/jira/browse/SOLR-9036> is relevant.

        Best,
        Erick

        On Thu, Nov 24, 2016 at 3:10 AM, Arkadi Colson
        <ark...@smartbit.be <mailto:ark...@smartbit.be>> wrote:

            This is the code from the master node. Al configs are the
            same on all nodes.
            I always add slaves with the collection API. Is there an
            other place to look
            for this part of the config?



            On 24-11-16 12:02, Michael Aleythe, Sternwald wrote:

                You need to change this on the master node. The part
                of the config you
                pasted here, looks like it is from the slave node.

                -----Ursprüngliche Nachricht-----
                Von: Arkadi Colson [mailto:ark...@smartbit.be
                <mailto:ark...@smartbit.be>]
                Gesendet: Donnerstag, 24. November 2016 11:56
                An: solr-user@lucene.apache.org
                <mailto:solr-user@lucene.apache.org>
                Betreff: Re: AW: Resync after restart

                Hi Michael

                Thanks for the quick response! The line does not exist
                in my config. So
                can I assume that the default configuration is to not
                replicate at startup?

                     <requestHandler name="/replication"
                class="solr.ReplicationHandler">
                       <lst name="defaults">
                         <str name="maxWriteMBPerSec">18.75</str>
                         <str name="commitReserveDuration">05:00:00</str>
                         <str name="httpConnTimeout">150000</str>
                         <str name="httpReadTimeout">300000</str>
                       </lst>
                     </requestHandler>

                Any other idea's?


                On 24-11-16 11:49, Michael Aleythe, Sternwald wrote:

                    Hi Arkadi,

                    you need to remove the line "<str
                    name="replicateAfter">startup</str>"
                    from your ReplicationHandler-config in
                    solrconfig.xml ->
                    https://wiki.apache.org/solr/SolrReplication
                    <https://wiki.apache.org/solr/SolrReplication>.

                    Greetings
                    Michael

                    -----Ursprüngliche Nachricht-----
                    Von: Arkadi Colson [mailto:ark...@smartbit.be
                    <mailto:ark...@smartbit.be>]
                    Gesendet: Donnerstag, 24. November 2016 09:26
                    An: solr-user <solr-user@lucene.apache.org
                    <mailto:solr-user@lucene.apache.org>>
                    Betreff: Resync after restart

                    Hi

                    Almost every time when restarting a solr instance
                    the index is replicated
                    completely. Is there a way to avoid this somehow?
                    The index currently has a
                    size of about 17GB.
                    Some advice here would be great.

                    99% of the config is defaul:

                    <updateLog> <str
                    name="dir">${solr.ulog.dir:}</str> <int
                    
name="numVersionBuckets">${solr.ulog.numVersionBuckets:65536}</int>
                    </updateLog> <autoCommit>
                               <maxTime>${solr.autoCommit.ma
                    <http://solr.autoCommit.ma>xTime:15000}</maxTime>
                               <openSearcher>false</openSearcher>
                             </autoCommit>

                    If you need more info, just let me know...

                    Thx!
                    Arkadi



Reply via email to