[
https://issues.apache.org/jira/browse/COUCHDB-1080?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Filipe Manana updated COUCHDB-1080:
-----------------------------------
Attachment: COUCHDB-1080-4-fdmanana.patch
Thanks Randall. It was well spotted this issue, which affects both replicators.
I made further changes to be even more explicit when there's an error updating
one of the checkpoints. It no longer bad matches when it fails to update a
checkpoint and it sends a message like the following back to the client:
{"error":"checkpoint_commit_failure","reason":"Error updating the source
checkpoint document: conflict"}
Also added the following clause:
+ {_NewSrcInstanceStartTime, _NewTgtInstanceStartTime} ->
+ {checkpoint_commit_failure, <<"Source and target databases out of "
+ "sync. Try to increase max_dbs_open at both servers.">>}
Maybe very unlikely to happen, but one never knows :)
I'll commit this 4th patch later today or tomorrow if no complaints are raised.
thanks again
> fail fast with checkpoint conflicts
> -----------------------------------
>
> Key: COUCHDB-1080
> URL: https://issues.apache.org/jira/browse/COUCHDB-1080
> Project: CouchDB
> Issue Type: Improvement
> Components: Replication
> Affects Versions: 1.0.2
> Reporter: Randall Leeds
> Fix For: 1.1, 1.2
>
> Attachments: COUCHDB-1080-2-fdmanana.patch,
> COUCHDB-1080-3-fdmanana.patch, COUCHDB-1080-4-fdmanana.patch,
> COUCHDB-1080-fdmanana.patch, paranoid_checkpoint_failure.patch,
> paranoid_checkpoint_failure_v2.patch
>
>
> I've thought about this long and hard and probably should have submitted the
> bug a long time ago. I've also run this in production for months.
> When a checkpoint conflict occurs it is almost always the right thing to do
> to abort.
> If there is a rev mismatch it could mean there's are two conflicting
> (continuous and one-shot) replications between the same hosts running.
> Without reloading the history documents checkpoints will continue to fail
> forever. This could leave us in a state with many replicated changes but no
> checkpoints.
> Similarly, a successful checkpoint but a lost/timed-out response could cause
> this situation.
> Since the supervisor will restart the replication anyway, I think it's safer
> to abort and retry.
--
This message is automatically generated by JIRA.
-
For more information on JIRA, see: http://www.atlassian.com/software/jira
