Varun Thacker created SOLR-12047: ------------------------------------ Summary: Solr 7.x restart can fail to load some cores Key: SOLR-12047 URL: https://issues.apache.org/jira/browse/SOLR-12047 Project: Solr Issue Type: Bug Security Level: Public (Default Security Level. Issues are Public) Affects Versions: 7.0 Reporter: Varun Thacker
I've seen this with 2 users running Solr 7.2.1 in the last 2 days where a restart fails to load some cores on a node. Here's the stack trace {noformat} date time ERROR (coreLoadExecutor-6-thread-2-processing-n:solr-number:8983_solr) [c:name s:shard r:core_node130 x:collection_shard_replica] o.a.s.c.ZkController org.apache.solr.common.SolrException: coreNodeName core_node130 does not exist in shard shard4: DocCollection(collection_name//collections/collection_name/state.json/2385)={ ..collection state.json ... } at org.apache.solr.cloud.ZkController.checkStateInZk(ZkController.java:1687) at org.apache.solr.cloud.ZkController.preRegister(ZkController.java:1590) at org.apache.solr.core.CoreContainer.createFromDescriptor(CoreContainer.java:1030) ... at java.lang.Thread.run(Thread.java:748) date time ERROR (coreContainerWorkExecutor-2-thread-1-processing-n:solr-number:8983_solr) [ ] o.a.s.c.CoreContainer Error waiting for SolrCore to be created java.util.concurrent.ExecutionException: org.apache.solr.common.SolrException: Unable to create core [collection_shardX_replica_n129] ... at java.lang.Thread.run(Thread.java:748) Caused by: org.apache.solr.common.SolrException: Unable to create core [collection_shardX_replica_n129] ... ... 5 more Caused by: org.apache.solr.common.SolrException: at org.apache.solr.cloud.ZkController.preRegister(ZkController.java:1619) at org.apache.solr.core.CoreContainer.createFromDescriptor(CoreContainer.java:1030) ... 7 more{noformat} I created the Jira saying Solr 7.x since it's tied to legacyCloud being set to false by default starting Solr 7.0 In ZkController#checkStateInZk where the block is only run with legacyCloud=false ( L1645 ) we do a waitForState ( L1667 ) and only wait 3 seconds. If we don't get the desired state the core will fail to load With big enough clusters this 3 second timeout is too low and we should increase it to a large number such that we don't cause core initialization failures Line reference is from Solr 7.2.1 -- This message was sent by Atlassian JIRA (v7.6.3#76005) --------------------------------------------------------------------- To unsubscribe, e-mail: dev-unsubscr...@lucene.apache.org For additional commands, e-mail: dev-h...@lucene.apache.org