Source and target can both be remote, that's not a problem for the replicator.
The addressinuse means couchdb's attempt to bind to port 5984 failed
because something else already bound it. Check to see if a beam.smp
process keeps launching and crashing. The 'badmatch' is that couchdb
gets that {error, eadd..} tuple instead of the expected chunk of a
reply.
B.
On 30 August 2013 10:01, Dave Cottlehuber <[email protected]> wrote:
> I'd more suspect
>
> {exit,eaddrinuse,
>
> which says that there's another process using the port. This is
> probably not obvious but you're asking your local couchdb to replicate
> between two remote instances this way -- both parameters are URLs --
> and I think that's why this is breaking:
>
> curl -H 'Content-Type: application/json' -X POST
> http://localhost:5984/_replicate -d \
> {"source":"http://isaacs.iriscouch.com/registry";, \
> "target":"http://admin:[email protected]:5984/registry"}'
>
> Try re-arranging this so that target is used as a local db rather than
> a "remote" url-based one:
>
> curl -vX PUT http://admin:passwd@localhost:5984/_replicator/npm_registry \
> -HContent-Type:application/json \
> --data-binary \
> '{"source":"http://isaacs.iriscouch.com/registry";,
> "target":"registry", "continuous": true, "user_ctx": {"name":
> "admin"}}'
>
> I also snuck in here to use the replicator db so this is persistent
> between reboots, and also to run under admin user context so it can
> also push ddoc updates as well.
>
> If that fails, other than that, what OS are you on, and are you sure
> there's nothing else (e.g. another couch instance) running?
>
>
> On 30 August 2013 09:14, Sebastian <[email protected]> wrote:
>> Hi,
>>
>> i am currently trying to replicate a couchDB Database from the net (in this
>> case the npmjs repo at http://isaacs.iriscouch.com/registry/) with the
>> command:
>>
>> curl -H 'Content-Type: application/json' -X POST
>> http://localhost:5984/_replicate -d '
>> {"source":"http://isaacs.iriscouch.com/registry","target":"http://admin:[email protected]:5984/registry"}'
>>
>> Unfortunately i am running in errors which happen a while after the
>> replication starts up:
>> [error] [emulator] Error in process <0.1337.0> with exit value:
>> {{badmatch,{"--cace6be2d1fb028f452235fec8c5879b",[]}},[{couch_httpd,split_header,1},{couch_httpd,'-parse_part_header/1-fun-1-',2},{lists,foldl,3},{couch_httpd,parse_part_header,1},{couch_httpd,parse_multipart_request...
>> [error] [<0.105.0>] {error_report,<0.30.0>,
>> {<0.105.0>,crash_report,
>> [[{initial_call,
>>
>> {mochiweb_socket_server,init,['Argument__1']}},
>> {pid,<0.105.0>},
>> {registered_name,[]},
>> {error_info,
>> {exit,eaddrinuse,
>> [{gen_server,init_it,6},
>> {proc_lib,init_p_do_apply,3}]}},
>> {ancestors,
>> [couch_secondary_services,couch_server_sup,
>> <0.31.0>]},
>> {messages,[]},
>> {links,[<0.93.0>]},
>> {dictionary,[]},
>> {trap_exit,true},
>> {status,running},
>> {heap_size,987},
>> {stack_size,24},
>> {reductions,459}],
>> []]}}
>>
>> I suppose it is because one of the entries is "bogus". After some search i
>> found these issues:
>> https://issues.apache.org/jira/browse/COUCHDB-1368
>> https://issues.apache.org/jira/browse/COUCHDB-1639
>>
>> which are supposed to be fixed in 1.3.1, which i am running and also
>> iriscouch has recently updated to.
>> Is there still an issue with this in 1.3.1 or am i missing something here?
>>
>> Sebastian
