[ https://issues.apache.org/jira/browse/AMQ-6453?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15664080#comment-15664080 ]
Xiang Zhao commented on AMQ-6453: --------------------------------- I think it might be the same issue as https://issues.apache.org/jira/browse/AMQ-5618 (which is still open after 21 months) or https://issues.apache.org/jira/browse/AMQ-6174 I'm using version 5.13.4, and had the similar issue: after fail over several times, I got the error "No reader available for position" or "Invalid log position: 0 " I suggest adding some warning on the related doc page, "ReplicatedLevelDB is not ready for production use". > (Replicated)LevelDB becomes corrupt when under stress > ----------------------------------------------------- > > Key: AMQ-6453 > URL: https://issues.apache.org/jira/browse/AMQ-6453 > Project: ActiveMQ > Issue Type: Bug > Components: activemq-leveldb-store > Affects Versions: 5.14.0 > Environment: OS: RHEL6; Virtual Machines (set of 3) > Reporter: Michiel Lange > > When running ActiveMQ 5.14.0 in a set of 3 brokers with ReplicatedLevelDB, > sending 100.000 messages with 100 threads (with jmeter) the persistency store > becomes corrupt (becoming evident when restarting a broker); Restarting a > broker with "1" replicas allows the broker to be started again and this seems > to work to allow to start the broker again -- This message was sent by Atlassian JIRA (v6.3.4#6332)