[ https://issues.apache.org/jira/browse/SOLR-1282?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12731493#action_12731493 ]
Brian Loudon commented on SOLR-1282: ------------------------------------ The dummy1/2 URLs are there only to more clearly show whether the slave attempted to use the configured or specified masterUrl for a particular request, as the slave generates an error message stating one or the other. The random failure to use the value of the masterUrl parameter still occurs when the slave does not have a masterUrl configured. In all cases the slave's response to the fetchindex command is 'OK', and any failure must be determined through examining logs. I added logging calls to the top of doSnapPull in org.apache.solr.handler.ReplicationHandler to display the contents of the SolrParams instance it is given. On requests where the masterUrl parameter appears to be ignored the logs show that the SolrParams instance contains no parameters. > Replication handler's fetchindex command randomly ignores masterUrl parameter > ----------------------------------------------------------------------------- > > Key: SOLR-1282 > URL: https://issues.apache.org/jira/browse/SOLR-1282 > Project: Solr > Issue Type: Bug > Components: replication (java) > Affects Versions: 1.4 > Reporter: Brian Loudon > Assignee: Noble Paul > > The fetchindex command of the replication handler is randomly ignoring the > value of a specified masterUrl parameter, instead using the masterUrl > configured in solrconfig.xml. > Injecting some debugging code shows that the contents of solrParams disappear > sometime after the thread running doSnapPull() starts in ReplicationHandler's > handleRequestBody() method. > This behavior is observed in the most recent nightly build (2009-07-14). > How to replicate: > Configure a core as a slave with a masterUrl: > <requestHandler name="/replication" class="solr.ReplicationHandler"> > <lst name="slave"> > <str name="masterUrl">http://dummy1</str> > </lst> > </requestHandler> > Run the core's fetchindex replication command several times, passing a > different masterUrl: > /replication?command=fetchindex&masterUrl=http://dummy2 > Logs show that the configured masterUrl (dummy1) is being used for at > least half the requests. -- This message is automatically generated by JIRA. - You can reply to this email to add a comment to the issue online.