> On Nov. 25, 2014, 7:34 p.m., Hrishikesh Gadre wrote: > > trunk/solr/core/src/java/org/apache/solr/core/ConfigSolrXml.java, line 55 > > <https://reviews.apache.org/r/28393/diff/1/?file=774307#file774307line55> > > > > This configuration property needs to correlate with the > > connection_timeout configuration specified in the servlet container hosting > > Solr. If Solr provides a default value, the user may not realize this (i.e. > > they may have a different connection_timeout configuration and if they use > > default value for this property, then they may experience *more* connection > > reset errors since the staleness check would be disabled). Is it possible > > to make this a required property? > > Gregory Chanan wrote: > Maybe it makes sense to do this with SOLR-4792 where we control > everything, i.e. we don't have to worry about another container's settings? > > Hrishikesh Gadre wrote: > In that case can we keep the previous settings in case this new parameter > is not specified (i.e. if this parameter is not specified, don't disable the > staleness check)? > > The advantage is that staless check would keep the failure window fairly > small (I have found it difficult to reproduce in a live cluster). With this > change, depending upon the difference between default_value (40) and the > actual connection_timeout configured, users may observe more failures (larger > the difference -> more failures). > > Hrishikesh Gadre wrote: > >>With this change, depending upon the difference between default_value > (40) and the actual connection_timeout configured, users may observe more > failures (larger the difference -> more failures). > > With the caveat that default_value > connection_timeout. > > Mark Miller wrote: > We will ship with a working default and there is a comment where you > would change alerting. I don't think it's a good idea to make it a required > property. It is the weakest part of this approach, but I don't see a better > option yet.
bq. Maybe it makes sense to do this with SOLR-4792 We are, this is 5.0 development work. - Mark ----------------------------------------------------------- This is an automatically generated e-mail. To reply, visit: https://reviews.apache.org/r/28393/#review63024 ----------------------------------------------------------- On Nov. 24, 2014, 3:42 p.m., Mark Miller wrote: > > ----------------------------------------------------------- > This is an automatically generated e-mail. To reply, visit: > https://reviews.apache.org/r/28393/ > ----------------------------------------------------------- > > (Updated Nov. 24, 2014, 3:42 p.m.) > > > Review request for lucene. > > > Repository: lucene > > > Description > ------- > > https://issues.apache.org/jira/browse/SOLR-4509 > > > Diffs > ----- > > trunk/solr/core/src/java/org/apache/solr/core/ConfigSolr.java 1641405 > trunk/solr/core/src/java/org/apache/solr/core/ConfigSolrXml.java 1641405 > trunk/solr/core/src/java/org/apache/solr/core/ConfigSolrXmlOld.java 1641405 > trunk/solr/core/src/java/org/apache/solr/core/CoreContainer.java 1641405 > trunk/solr/core/src/java/org/apache/solr/core/PluginInfo.java 1641405 > > trunk/solr/core/src/java/org/apache/solr/handler/component/HttpShardHandlerFactory.java > 1641405 > trunk/solr/core/src/java/org/apache/solr/update/UpdateShardHandler.java > 1641405 > > trunk/solr/core/src/java/org/apache/solr/util/IdleConnectionMonitorRunnable.java > PRE-CREATION > trunk/solr/core/src/test/org/apache/solr/cloud/OverseerTest.java 1641405 > > trunk/solr/core/src/test/org/apache/solr/cloud/TestLeaderElectionZkExpiry.java > 1641405 > trunk/solr/core/src/test/org/apache/solr/cloud/TestZkChroot.java 1641405 > trunk/solr/core/src/test/org/apache/solr/cloud/ZkControllerTest.java > 1641405 > trunk/solr/core/src/test/org/apache/solr/core/TestCoreDiscovery.java > 1641405 > > trunk/solr/core/src/test/org/apache/solr/core/TestImplicitCoreProperties.java > 1641405 > trunk/solr/server/etc/jetty.xml 1641405 > > trunk/solr/solrj/src/java/org/apache/solr/client/solrj/impl/HttpClientUtil.java > 1641405 > > Diff: https://reviews.apache.org/r/28393/diff/ > > > Testing > ------- > > > Thanks, > > Mark Miller > >
