Uploaded image for project: 'Sakai'
  1. Sakai
  2. SAK-27671

"node already merging shared segments" on lone instance

    Details

    • Type: Bug Bug
    • Status: Resolved
    • Priority: Blocker Blocker
    • Resolution: Won't Fix
    • Affects Version/s: 2.7.2, 2.8.0, 2.8.1
    • Fix Version/s: None
    • Component/s: Search
    • Labels:
    • Previous Issue Keys:
      SAK-16947, SRCH-38

      Description

      I have a lone instance which prints the following error into the log every 10 seconds:

      2009-09-09 11:02:27,502 ERROR Timer-4 org.sakaiproject.search.optimize.shared.impl.DbJournalOptimizationManager - This node already merging shared segments, index writer spellbreaker:1251938321559 This node is currently optimizing the shared segments, This is an error as only one copy of this node should be Active in the cluster

      In spite of the message, this instance has never been part of a cluster.

        Gliffy Diagrams

          Issue Links

            Activity

            Hide
            Swe Aung added a comment - - edited

            Hi
            I am facing the same error after restarting the nodes. I tried running above SQL command, but the error did not stop. I am running 2.8.0 with 2 nodes setup with Oracle 10g database.
            Please advise. Your prompt reply is appreciated.

            Show
            Swe Aung added a comment - - edited Hi I am facing the same error after restarting the nodes. I tried running above SQL command, but the error did not stop. I am running 2.8.0 with 2 nodes setup with Oracle 10g database. Please advise. Your prompt reply is appreciated.
            Hide
            John Bush added a comment - - edited

            This will not be an issue with the elasticsearch impl the whole way the storage of index data is managed is completely different. The recommended solution to addressing these sort of issues is to switch over to elasticsearch in Sakai 10. These types of issues is the whole reason search was re-implemented.

            Show
            John Bush added a comment - - edited This will not be an issue with the elasticsearch impl the whole way the storage of index data is managed is completely different. The recommended solution to addressing these sort of issues is to switch over to elasticsearch in Sakai 10. These types of issues is the whole reason search was re-implemented.
            Hide
            Matthew Jones added a comment -

            Is elasticsearch the default in trunk? How is it being activated/switched from the lucene search? It looks like the parallelIndexComponents.xml and the optimizer in that are still active and I'm still seeing this error running of recent trunk build. Do you have any examples of avoiding this, how to use SRCH-119 or why this would still be happening?

            Show
            Matthew Jones added a comment - Is elasticsearch the default in trunk? How is it being activated/switched from the lucene search? It looks like the parallelIndexComponents.xml and the optimizer in that are still active and I'm still seeing this error running of recent trunk build. Do you have any examples of avoiding this, how to use SRCH-119 or why this would still be happening?
            Hide
            John Bush added a comment -

            I don't think elasticsearch is the default in trunk right now, we've gone back and forth on that. I think it should be the default, b/c while the solr option might be just as good (i have no experience with it) it does require another server be setup. To enable elasticsearch as the default the following should be set:

            search.enable=true
            elasticsearch.http.enabled=true
            elasticsearch.http.port=9200
            search.service.impl=org.sakaiproject.search.elasticsearch.ElasticSearchService
            search.indexbuilder.impl=org.sakaiproject.search.elasticsearch.ElasticSearchIndexBuilder

            Show
            John Bush added a comment - I don't think elasticsearch is the default in trunk right now, we've gone back and forth on that. I think it should be the default, b/c while the solr option might be just as good (i have no experience with it) it does require another server be setup. To enable elasticsearch as the default the following should be set: search.enable=true elasticsearch.http.enabled=true elasticsearch.http.port=9200 search.service.impl=org.sakaiproject.search.elasticsearch.ElasticSearchService search.indexbuilder.impl=org.sakaiproject.search.elasticsearch.ElasticSearchIndexBuilder
            Hide
            Matthew Jones added a comment -

            Elasticsearch is what's in there out of the box, certainly lucene shouldn't be the default for 10 and SOLR requires some configuration that doesn't exist. Thanks for the info. I can create a new jira to make ElasticSearchService the default too..

            Show
            Matthew Jones added a comment - Elasticsearch is what's in there out of the box, certainly lucene shouldn't be the default for 10 and SOLR requires some configuration that doesn't exist. Thanks for the info. I can create a new jira to make ElasticSearchService the default too..

              People

              • Assignee:
                John Bush
                Reporter:
                Joshua Swink
              • Votes:
                0 Vote for this issue
                Watchers:
                11 Start watching this issue

                Dates

                • Created:
                  Updated:
                  Resolved:

                  Development

                    Git Source Code