[ https://issues.apache.org/jira/browse/HBASE-18459?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Enis Soztutar updated HBASE-18459: ---------------------------------- Attachment: hbase-18459-v1.patch v1 patch. Formatting is off (but I'll fix it in commit). I've tested this manually with load-client with 10 threads, and shutting down the zookeeper before the client starts. > [C++] Fix Segfault in location-cache > ------------------------------------ > > Key: HBASE-18459 > URL: https://issues.apache.org/jira/browse/HBASE-18459 > Project: HBase > Issue Type: Sub-task > Reporter: Enis Soztutar > Assignee: Enis Soztutar > Fix For: HBASE-14850 > > Attachments: hbase-18459-v1.patch > > > When testing with the new multi-threaded load-client (implemented in > HBASE-18086), I've noticed that sometimes we get a segfault when zookeeper or > master connection gets an exception. > It seems to be from the way the we invalidate meta and refresh the meta. -- This message was sent by Atlassian JIRA (v6.4.14#64029)