[ https://issues.apache.org/jira/browse/HBASE-20576?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16475268#comment-16475268 ]
Zheng Hu commented on HBASE-20576: ---------------------------------- {code} + if (!remoteWALDirForPeer.getFileSystem(env.getMasterConfiguration()) + .exists(remoteWALDirForPeer)) { + throw new DoNotRetryIOException( + "The remote WAL directory " + remoteWALDirForPeer.toUri().toString() + " does not exist"); + } {code} If pass an invalid path which has no schema (such as, REMOTE_WAL_DIR=>'/hbase/remoteWALs' ), the remoteWALDirForPeer.getFileSystem will use the default Fs in env.getMasterConfiguration(), which is incorrect ... Better to check the path when both peer adding & transiting. BTW, can just use remoteWALDirForPeer , instead of remoteWALDirForPeer.toUri().toString() in {code} + throw new DoNotRetryIOException( + "The remote WAL directory " + remoteWALDirForPeer.toUri().toString() + " does not exist"); {code} Here the SupressWarning annotation could be removed ? {code} + @SuppressWarnings("unchecked") + private void replayRemoteWAL() { + addChildProcedure(new RecoverStandbyProcedure(peerId)); + } {code} > Check whether the remote WAL directory is provided when transiting to A > ----------------------------------------------------------------------- > > Key: HBASE-20576 > URL: https://issues.apache.org/jira/browse/HBASE-20576 > Project: HBase > Issue Type: Sub-task > Components: Replication > Reporter: Duo Zhang > Assignee: Duo Zhang > Priority: Major > Fix For: HBASE-19064 > > Attachments: HBASE-20576-HBASE-19064-v1.patch, > HBASE-20576-HBASE-19064-v1.patch, HBASE-20576-HBASE-19064-v1.patch, > HBASE-20576-HBASE-19064.patch, HBASE-20576-HBASE-19064.patch > > > When testing on the real cluster I typed a wrong remote wal directory. Then I > start the procedure for transiting the peer to A, the procedure was stuck > there for ever... -- This message was sent by Atlassian JIRA (v7.6.3#76005)