[ https://issues.apache.org/jira/browse/SOLR-2323?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Alexandre Rafalovitch closed SOLR-2323. --------------------------------------- Resolution: Won't Fix An issue against the replication method no longer used in Solr. > Solr should clean old replication temp dirs > ------------------------------------------- > > Key: SOLR-2323 > URL: https://issues.apache.org/jira/browse/SOLR-2323 > Project: Solr > Issue Type: Bug > Components: replication (java) > Affects Versions: 1.4.1 > Reporter: Markus Jelsma > > In a high commit rate environment (polling < 10s and commits every minute) > the shutdown/restart of a slave can result in old temp directories laying > around, filling up disk space as we go on. This happens with the following > scenario: > 1. master has index version 2 > 2. slave downloads files for version 2 to index.2 temp directory > 3. slave is shutdown > 4. master increments to version 3 > 5. slave is started > 6. slave downloads files for version 3 to index.3 temp directory > The result is index.2 temp directory not getting deleted by any process. This > is very annoying in such an environment where nodes are restarted frequently > (for whatever reason). Working around the problem can be done by either > manually deleting the temp directories between shutdown and startup or by > calling the disablepoll command followed by an abortfetch command which will > (after a long wait) finally purge the temp directory. > See this thread: > http://www.mail-archive.com/solr-user@lucene.apache.org/msg45120.html -- This message was sent by Atlassian JIRA (v6.3.4#6332) --------------------------------------------------------------------- To unsubscribe, e-mail: dev-unsubscr...@lucene.apache.org For additional commands, e-mail: dev-h...@lucene.apache.org