[ https://issues.apache.org/jira/browse/ZOOKEEPER-1364?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13191402#comment-13191402 ]
Patrick Hunt commented on ZOOKEEPER-1364: ----------------------------------------- See this prior related discussion: http://markmail.org/message/ysymrofslp2opqei A while back I had used aspectj to do a one-off for this. Lost track of it since. However the basic idea was to using pointcuts on the network/filesystem read/write operations in order to introduce failures. Worked well (for example the quorum would be lost and then re-established, much error handling code was exercised. I had to monitor it by hand though and ensure that things were functioning as expected. I found at least 3 or 4 issues with this approach. > Add orthogonal fault injection mechanism/framework > -------------------------------------------------- > > Key: ZOOKEEPER-1364 > URL: https://issues.apache.org/jira/browse/ZOOKEEPER-1364 > Project: ZooKeeper > Issue Type: Test > Components: tests > Reporter: Andrei Savu > Assignee: Andrei Savu > > Hadoop has a mechanism for doing fault injection (HDFS-435). I think it would > be useful if something similar would be available for ZooKeeper. -- This message is automatically generated by JIRA. If you think it was sent incorrectly, please contact your JIRA administrators: https://issues.apache.org/jira/secure/ContactAdministrators!default.jspa For more information on JIRA, see: http://www.atlassian.com/software/jira