"Object Not Found" error while pullReplication

51 views
Skip to first unread message

Azhar Bandri

unread,
Sep 17, 2014, 2:12:11 AM9/17/14
to mobile-c...@googlegroups.com
I have a Android application which does only pull replication from CouchDB. The code and replication was working fine for about 2 months and suddenly it stopped replicating and it gives following error in log cat, 
E/RemoteRequest(17030): Got error status: 404 for http://ip:port/DBName/_local/22ba996d673e136e09af3d780021718fda7a83c5.  Reason: Object Not Found

one more thing I noticed is "22ba996d673e136e09af3d780021718fda7a83c5" is 40 digit while my actual document Ids consist of 32 digit.

Any Idea what must be wrong here?

Jens Alfke

unread,
Sep 17, 2014, 11:48:59 AM9/17/14
to mobile-c...@googlegroups.com

On Sep 16, 2014, at 11:12 PM, Azhar Bandri <azhar....@gmail.com> wrote:

I have a Android application which does only pull replication from CouchDB. The code and replication was working fine for about 2 months and suddenly it stopped replicating and it gives following error in log cat, 
E/RemoteRequest(17030): Got error status: 404 for http://ip:port/DBName/_local/22ba996d673e136e09af3d780021718fda7a83c5.  Reason: Object Not Found

That's not actually an error; the Android CBL shouldn't be logging it. Those "_local/…" documents are checkpoints that are used to save replication state. Not finding a checkpoint just means that the replicator will re-build the replication state, which is a one-time slowdown.

If replication is failing, there must be some other reason behind it. What other symptoms or logs do you see? (And at this point I'll probably have to hand the phone over to Traun, because I don't work on the Android version.)

—Jens

Azhar Bandri

unread,
Sep 18, 2014, 1:07:55 AM9/18/14
to mobile-c...@googlegroups.com

If replication is failing, there must be some other reason behind it. What other symptoms or logs do you see? 

—Jens

Below the Error the Log shows these,

W/Sync(10089): Puller@421d3bb0[http://ip:port/DBName/]: starting ChangeTracker with since=null mode=OneShot
W/Sync(10089): Puller@421d3bb0[http://ip:port/DBName/]: started ChangeTracker com.couchbase.lite.replicator.ChangeTracker@4227baa0
 

Traun Leyden

unread,
Sep 18, 2014, 2:09:46 PM9/18/14
to mobile-c...@googlegroups.com

One good way to debug this is to use a tool like HTTPScoop or CharlesProxy.  (or an even lower level tool like ngrep or tcpdump)

I wrote a blog article about how to use CharlesProxy, and HTTPScoop is even easier to use but I don't have a writeup for it.  (essentially you just need to make sure it is scooping on the ports you are using, typically port 4984 or port 80)

With those tools you can see the traffic between the replicator and sync gateway, and you can see if the replicator is active or not.  For example the pull replicator should be constantly hitting the _changes endpoint on the sync gateway.




--
You received this message because you are subscribed to the Google Groups "Couchbase Mobile" group.
To unsubscribe from this group and stop receiving emails from it, send an email to mobile-couchba...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/mobile-couchbase/f550699e-1167-4572-8e16-3ad2a55c0fd5%40googlegroups.com.

For more options, visit https://groups.google.com/d/optout.

Reply all
Reply to author
Forward
0 new messages