Mailing-List: contact user-help@couchdb.apache.org; run by ezmlm
Precedence: bulk
Reply-To: user@couchdb.apache.org
Received-SPF: pass (nike.apache.org: local policy)
Message-Id: 
 <1366707924.31855.140661221524722.1A7F1BA9@webmail.messagingengine.com>
From: Calle Arnesten <calle.arnesten@codekick.com>
To: user@couchdb.apache.org
MIME-Version: 1.0
Content-Transfer-Encoding: 7bit
Content-Type: text/plain
Subject: Re: Problem with restart of continuous replication
Date: Tue, 23 Apr 2013 11:05:24 +0200

Thanks for your answer.

I was using persistent _replicator. It was set up like this from both ends:
curl -X PUT -d "{ \"source\":\"<LOCAL_DB>\", \"target\":\"<REMOTE_DB>\", \"continuous\":true }"\
                 <REMOTE>/_replicator/<A_NAME> -H "Content-Type:application/json"

The strange thing was that after the 2 hour communication problem, the replication was automatically restarted on one of the ways but not the other way. 

/Calle


On Tue, Apr 23, 2013, at 10:15, Robert Newson wrote:
> I recommend using the persistent _replicator replication manager, it
> will restart your replication jobs after a crash. POST's to _replicate
> with continuous:true will continue until a) the connection is severed
> or b) the node reboots, whichever comes first.
> 
> http://wiki.apache.org/couchdb/Replication#Replicator_database
> 
> B.
> 
> On 23 April 2013 03:01, Calle Arnesten <calle.arnesten@codekick.com> wrote:
> > Hi,
> >
> > I have a two-way continuous replication set up between two servers with CouchDB. About a week ago we had a DNS problem that caused the machines to not to be able to communicate for almost 2 hours. My assumption was that CouchDB would automatically try to restart the replication. When I double-checked this a day or two later, replication was only restarted for one way and not both. After rebooting the machine the replication started working both ways again.
> >
> > Now my question is: Was my assumption wrong that replication should be restarted automatically? If not, what could be the cause of it not working? The machine that it happened to have CouchDB 1.2.1 installed.
> >
> > Also, do you have any recommendations on how to monitor that replication works and notify by mail when it doesn't?
> >
> > /Calle