[jira] [Comment Edited] (HBASE-17290) Potential loss of data for replication of bulk loaded hfiles
[ https://issues.apache.org/jira/browse/HBASE-17290?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15801226#comment-15801226 ] Ashish Singhi edited comment on HBASE-17290 at 1/5/17 12:25 PM: Sorry for the delay, got stuck with company work. I have attached the patch. Added a new RS observer, ReplicationObserver to solve this bug. Please review. was (Author: ashish singhi): Sorry, got stuck with company work. I have attached the patch. Added a new RS observer, ReplicationObserver to solve this bug. Please review. > Potential loss of data for replication of bulk loaded hfiles > > > Key: HBASE-17290 > URL: https://issues.apache.org/jira/browse/HBASE-17290 > Project: HBase > Issue Type: Bug >Affects Versions: 1.3.0 >Reporter: Ted Yu > Fix For: 2.0.0, 1.4.0 > > Attachments: HBASE-17290.patch > > > Currently the support for replication of bulk loaded hfiles relies on bulk > load marker written in the WAL. > The move of bulk loaded hfile(s) (into region directory) may succeed but the > write of bulk load marker may fail. > This means that although bulk loaded hfile is being served in source cluster, > the replication wouldn't happen. > Normally operator is supposed to retry the bulk load. But relying on human > retry is not robust solution. -- This message was sent by Atlassian JIRA (v6.3.4#6332)
[jira] [Comment Edited] (HBASE-17290) Potential loss of data for replication of bulk loaded hfiles
[ https://issues.apache.org/jira/browse/HBASE-17290?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15760715#comment-15760715 ] Ted Yu edited comment on HBASE-17290 at 12/27/16 2:57 PM: -- The latest patch for HBASE-14417 is on reviewboard and attached to JIRA. was (Author: yuzhih...@gmail.com): The latest patch for HBASE-14417 is on reviewboard. > Potential loss of data for replication of bulk loaded hfiles > > > Key: HBASE-17290 > URL: https://issues.apache.org/jira/browse/HBASE-17290 > Project: HBase > Issue Type: Bug >Reporter: Ted Yu > > Currently the support for replication of bulk loaded hfiles relies on bulk > load marker written in the WAL. > The move of bulk loaded hfile(s) (into region directory) may succeed but the > write of bulk load marker may fail. > This means that although bulk loaded hfile is being served in source cluster, > the replication wouldn't happen. > Normally operator is supposed to retry the bulk load. But relying on human > retry is not robust solution. -- This message was sent by Atlassian JIRA (v6.3.4#6332)
[jira] [Comment Edited] (HBASE-17290) Potential loss of data for replication of bulk loaded hfiles
[ https://issues.apache.org/jira/browse/HBASE-17290?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=15760710#comment-15760710 ] Ted Yu edited comment on HBASE-17290 at 12/19/16 3:46 PM: -- HBASE-15446 is still open. As the description of HBASE-15446 says, the solution depends on client to retry. What if client doesn't retry ? See the latest patch on HBASE-14417 for solution - we record the hfile before the movement into region directory using the hook introduced by HBASE-17292. was (Author: yuzhih...@gmail.com): HBASE-15446 is still open. As the description of HBASE-15446 says, the solution depends on client to retry. What if client doesn't retry ? See the latest patch on HBASE-14417 for solution - we record the hfile before the movement into region directory. > Potential loss of data for replication of bulk loaded hfiles > > > Key: HBASE-17290 > URL: https://issues.apache.org/jira/browse/HBASE-17290 > Project: HBase > Issue Type: Bug >Reporter: Ted Yu > > Currently the support for replication of bulk loaded hfiles relies on bulk > load marker written in the WAL. > The move of bulk loaded hfile(s) (into region directory) may succeed but the > write of bulk load marker may fail. > This means that although bulk loaded hfile is being served in source cluster, > the replication wouldn't happen. > Normally operator is supposed to retry the bulk load. But relying on human > retry is not robust solution. -- This message was sent by Atlassian JIRA (v6.3.4#6332)