couchdb-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Kevin R. Coombes" <kevin.r.coom...@gmail.com>
Subject Re: Replication of design docs fails on 1.0.1 (was Re: upgrade on windows)
Date Tue, 03 Jan 2012 20:12:12 GMT
I have good news and bad news.

The Good News:   I successfully upgraded to 1.0.2.  All of my JavaScript 
code still appears to work.

The Bad News: Pull replication from a remote machine to the local 
machine still fails on the design document -- maybe.  Here is a 
transcript of the new session

############
C:\MyStuff\MyCodeProjects\CouchReplication>type drugbase-from-gani-to-here
{
   "target" : "http://testonly:deleteme@bcbwfl6sjn1:5984/drugbase",
   "source" : "http://10.105.43.35:5984/drugbase",
   "user_ctx": {
      "name": "testonly",
      "roles": ["_admin"]
   }
}
C:\MyStuff\MyCodeProjects\CouchReplication>curl
  -H "Content-Type: application/json" -X POST -d @drugbase-from-gani-to-here
  http://testonly:deleteme@bcbwfl6sjn1:5984/_replicate
{"error":"json_encode","reason":"{bad_term,{couch_doc,fold_streamed_data,\n
                 [#Fun<couch_rep_att.0.90652905>,-1140,\n
#Fun<couch_doc.13.28980337>,ok]}}"}
############

And here is the relevant portion of the error log:

############
[Tue, 03 Jan 2012 19:43:43 GMT] [error] [<0.619.0>] streaming att. ended but more data
requested {1325,619823,339311}

[Tue, 03 Jan 2012 19:43:43 GMT] [info] [<0.619.0>] request for "_design%2Fbasic/docs.html"
timed out, retrying in 0.5 seconds

[Tue, 03 Jan 2012 19:43:43 GMT] [info] [<0.621.0>] 10.105.35.136 - - 'POST' /drugbase/_ensure_full_commit
201

[Tue, 03 Jan 2012 19:43:43 GMT] [info] [<0.572.0>] recording a checkpoint for http://10.105.43.35:5984/drugbase/
->  http://testonly:*****@bcbwfl6sjn1:5984/drugbase/ at source update_seq 16479

[Tue, 03 Jan 2012 19:43:43 GMT] [info] [<0.649.0>] 10.105.35.136 - - 'PUT' /drugbase/_local%2Fa4413563817bd705ab727a8ab1643a28
201

[Tue, 03 Jan 2012 19:43:43 GMT] [error] [emulator] Error in process<0.619.0>  with exit
value: {function_clause,[{couch_doc,fold_streamed_data,[#Fun<couch_rep_att.0.90652905>,-1140,#Fun<couch_doc.13.28980337>,ok]},{couch_doc,atts_to_mp,4},{couch_rep_writer,streamer_fun,3}]}
############

I will note that the error log has one more entry that intervenes, 
namely the 'PUT' to the /_local database.  I do't know how relevant that is.

Thanks again.
     Kevin

On 1/3/2012 12:22 PM, Dave Cottlehuber wrote:
>> On 3 January 2012 18:45, Kevin R. Coombes<kevin.r.coombes@gmail.com>  wrote:
>>> Thanks for all the advice.  I haven't upgraded yet; instead, inspired by the
>>> claim that "it works for me", I decided to poke around some more into the
>>> replication issue. I'm now going to report that replication of design docs
>>> *sometimes* fails on version 1.0.1, which is probably the worse kind of bug
>>> to report....
>>>
>>> First, here is a transcript of the (64-bit Windows 7) cmd window:
>>>
>>> #############
>>> C:\MyStuff\MyCodeProjects\CouchReplication>curl
>>> http://bcbwfl6sjn1:5984/testrepl/_all_docs
>>> {"total_rows":2,"offset":0,"rows":[
>>> {"id":"43567c4d142e9d03f6e966a50400089b","key":"43567c4d142e9d03f6e966a50400089b",
>>> "value":{"rev":"1-711a20fc5975380c99726b4da845d277"}},
>>> {"id":"_design/basic","key":"_design/basic","value":{"rev":
>>> "1-dddcf8b46c73b0102f86047765ab9284"}}
>>> ]}
>>>
>>> C:\MyStuff\MyCodeProjects\CouchReplication>curl
>>> http://bcbwfl6sjn1:5984/reptarget/_all_docs
>>> {"total_rows":0,"offset":0,"rows":[]}
>>>
>>> C:\MyStuff\MyCodeProjects\CouchReplication>type testrepl.json
>>> {
>>>   "target" : "http://testonly:deleteme@bcbwfl6sjn1:5984/reptarget",
>>>   "source" : "http://bcbwfl6sjn1:5984/testrepl",
>>>   "user_ctx": {
>>>      "name": "testonly",
>>>      "roles": ["_admin"]
>>>   }
>>> }
>>>
>>> C:\MyStuff\MyCodeProjects\CouchReplication>curl
>>>   -H "Content-Type: application/json" -X POST -d @testrepl.json
>>>   http://testonly:deleteme@bcbwfl6sjn1:5984/_replicate
>>> {"ok":true,"session_id":"d432de689e42bf1a983b8640b6835794","source_last_seq":2,
>>> "history":[{"session_id":"d432de689e42bf1a983b8640b6835794","start_time":
>>> "Tue, 03 Jan 2012 17:22:52 GMT","end_time":"Tue, 03 Jan 2012 17:22:52 GMT",
>>> "start_last_seq":0,"end_last_seq":2,"recorded_seq":2,"missing_checked":0,
>>> "missing_found":2,"docs_read":2,"docs_written":2,"doc_write_failures":0}]}
>>>
>>> C:\MyStuff\MyCodeProjects\CouchReplication>type drugbase-from-gani-to-here
>>> {
>>>   "target" : "http://testonly:deleteme@bcbwfl6sjn1:5984/drugbase",
>>>   "source" : "http://10.105.43.35:5984/drugbase",
>>>   "user_ctx": {
>>>      "name": "testonly",
>>>      "roles": ["_admin"]
>>>   }
>>> }
>>>
>>> C:\MyStuff\MyCodeProjects\CouchReplication>curl
>>>   -H "Content-Type: application/json" -X POST -d @drugbase-from-gani-to-here
>>>   http://testonly:deleteme@bcbwfl6sjn1:5984/_replicate
>>> {"error":"json_encode","reason":"{bad_term,{couch_doc,fold_streamed_data,\n
>>>                 [#Fun<couch_rep_att.0.90652905>,-1140,\n
>>> #Fun<couch_doc.13.28980337>,ok]}}"}
>>> #############
>>>
>>> Note that the first replication attempt (which works) goes from one database
>>> to another on the same machine. This attempt uses a small test database that
>>> has the usual sorts of design document attachments for a CouchApp along with
>>> only one ordinary document.
>>>
>>> The second attempt is a pull replication from a remote machine to the local
>>> machine.  This has the same kind of CouchApp design documents, and has about
>>> 16000 ordinary documents. All 16000+ ordinary documents get replicated
>>> without a problem.  Only yte design document fails.  Here is the start of
>>> the error message from the log file,to confirm that it is (an attachment to)
>>> the design document that fails:
>>>
>>> #############
>>> [Tue, 03 Jan 2012 17:25:25 GMT] [error] [<0.650.0>] streaming att. ended
but
>>> more data requested {1325,611525,197841}
>>>
>>> [Tue, 03 Jan 2012 17:25:25 GMT] [info] [<0.650.0>] request for
>>> "_design%2Fbasic/docs.html" timed out, retrying in 0.5 seconds
>>>
>>> [Tue, 03 Jan 2012 17:25:25 GMT] [error] [<0.650.0>] streaming att. ended
but
>>> more data requested {1325,611525,712841}
>>>
>>> [Tue, 03 Jan 2012 17:25:25 GMT] [info] [<0.650.0>] request for
>>> "_design%2Fbasic/docs.html" timed out, retrying in 0.5 seconds
>>>
>>> [Tue, 03 Jan 2012 17:25:26 GMT] [error] [emulator] Error in process
>>> <0.650.0>  with exit value:
>>> {function_clause,[{couch_doc,fold_streamed_data,[#Fun<couch_rep_att.0.90652905>,-1140,#Fun<couch_doc.13.28980337>,ok]},{couch_doc,atts_to_mp,4},{couch_rep_writer,streamer_fun,3}]}
>>> #############
>>>
>>> Any suggestions? (I am open to the suggestion that "upgrading to Version X
>>> will solve the problem" for a specific value of X....)
>>>
>>> Thanks,
>>>     Kevin
>>>
> Hi Kevin,
>
> Thanks, those logs helped a lot. There are a number of fixes that
> could be related;
>
> https://issues.apache.org/jira/browse/COUCHDB-868 and
> https://issues.apache.org/jira/browse/COUCHDB-1021 which if I read my
> repo history right, are both in 1.0.2.
>
> I suggest you try that first which might avoid you needing to update
> your JS stuff, and keep us posted.
>
> A+
> Dave

Mime
View raw message