Replication error : connection timeout #3957

Closed
marob opened this Issue Oct 1, 2013 · 3 comments

Comments

Projects
None yet
2 participants

marob commented Oct 1, 2013

Hi,

The replication was OK for several weeks/months on my environment and it failed suddenly this morning at 07:00 GMT (no known changes on the environment):

[Tue, 01 Oct 2013 06:58:57 GMT] [info] [<0.26816.20>] 127.0.0.1 - - GET /_replicator/npmjs.org 200
[Tue, 01 Oct 2013 06:59:42 GMT] [error] [<0.125.0>] Replicator, request GET to "http://isaacs.iriscouch.com/registry/" failed due to error {error,{conn_failed,{error,etimedout}}}
[Tue, 01 Oct 2013 06:59:42 GMT] [error] [<0.125.0>] ChangesReader process died with reason: {http_request_failed,
                                                         "GET",
                                                         "http://isaacs.iriscouch.com/registry/_changes?feed=continuous&style=all_docs&since=700278&heartbeat=10000",
                                                         {error,
                                                          {error,
                                                           {conn_failed,
                                                            {error,
                                                             etimedout}}}}}
[Tue, 01 Oct 2013 06:59:42 GMT] [error] [<0.125.0>] Replication `71c5aa7abb0d09e9410dc31d2f917cb1+continuous` (`http://isaacs.iriscouch.com/registry/` -> `registry`) failed: changes_reader_died
[Tue, 01 Oct 2013 06:59:42 GMT] [error] [<0.112.0>] Error in replication `71c5aa7abb0d09e9410dc31d2f917cb1+continuous` (triggered by document `npmjs.org`): changes_reader_died
Restarting replication in 5 seconds.

If I try to access http://isaacs.iriscouch.com/registry/_changes?feed=continuous&style=all_docs&since=700278&heartbeat=10000 from the same environment, the download of the changes takes forever (less than 1kB/s !).
Same problem on an other environment.
Perhaps this is why the connection times-out.

The replication is currently blocked on "source sequence 700549" while the current source sequence is 701512.

Is there a known problem on isaacs.iriscouch.com since this morning ?

Thanks

marob commented Oct 1, 2013

It seams to be back to normal...

Contributor

luk- commented Oct 3, 2013

Great :)

@luk- luk- closed this Oct 3, 2013

Contributor

luk- commented Oct 3, 2013

Also, guessing this was just a transient network/system issue.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment