couchdb-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Filipe David Manana <fdman...@apache.org>
Subject Re: Replicator exceptions
Date Sat, 26 Mar 2011 21:46:04 GMT
On Sat, Mar 26, 2011 at 5:39 PM, kowsik <kowsik@gmail.com> wrote:
> Filipe,
> Sorry for the late response. I'm seeing this in our production couch's
> and so will try to patch one of them to see if it works. Is there a
> way you can explicitly trigger this bug so we can write a unit test
> for it?

I just reproduced it now by doing the following:
1) create a database, add some docs and a design doc with a filter;
2) GET /db/_changes?feed=continuous&filter=ddocname/filtername
3) compact database, wait for compaction to finish
4) write 1 doc to the database - after this the _changes connection is
lost and you get a small stack trace in the log

I confirm that the patch fixes this particular scenario for trunk at
least. There might be others. Filtered replications will get they're
connection to _changes closed as well.


>
> Thanks,
>
> K
> ---
> http://blitz.io
> http://twitter.com/pcapr
>
> On Wed, Mar 23, 2011 at 7:22 PM, Filipe David Manana
> <fdmanana@apache.org> wrote:
>> Hi Kowsik,
>>
>> It's likely to be related to ticket 1093.
>> What you describe can be related to your replication issues if they
>> are filtered replications.
>> However the stack trace you've pasted is related to the view indexer.
>>
>> Any more useful stack traces in the log?
>>
>> Btw, I've just attached a patch to 1093. It would be great if you
>> could confirm it fixes your issue.
>>
>> regards,
>>
>> On Mon, Mar 21, 2011 at 1:50 PM, kowsik <kowsik@gmail.com> wrote:
>>> Saw this in the logs last night. I'm guessing this is similar to #1093
>>> but causes the replicator to quit. I have bidirectional replication
>>> (1.0.2) with _changes feed and periodic compaction. The last operation
>>> before this happened was _compact.
>>>
>>> ** Last message in was {'$gen_cast',reopen_db}
>>>
>>> [Tue, 08 Feb 2011 06:03:11 GMT] [error] [<0.3541.0>] {error_report,<0.31.0>,
>>>    {<0.3541.0>,crash_report,
>>>     [[{initial_call,{couch_view_group,init,['Argument__1']}},
>>>       {pid,<0.3541.0>},
>>>       {registered_name,[]},
>>>       {error_info,
>>>           {exit,
>>>               {noproc,{gen_server,call,[<0.3498.0>,{drop,<0.3541.0>}]}},
>>>               [{gen_server,terminate,6},{proc_lib,init_p_do_apply,3}]}},
>>>       {ancestors,
>>>           [couch_view,couch_secondary_services,couch_server_sup,<0.32.0>]},
>>>       {messages,[]},
>>>       {links,[<0.3546.0>,<0.3548.0>,<0.101.0>]},
>>>       {dictionary,[]},
>>>       {trap_exit,true},
>>>       {status,running},
>>>       {heap_size,2584},
>>>       {stack_size,24},
>>>       {reductions,2009}],
>>>      [{neighbour,
>>>           [{pid,<0.3546.0>},
>>>            {registered_name,[]},
>>>            {initial_call,{couch_event_sup,init,['Argument__1']}},
>>>            {current_function,{gen_server,loop,6}},
>>>            {ancestors,
>>>                [<0.3541.0>,couch_view,couch_secondary_services,
>>>                 couch_server_sup,<0.32.0>]},
>>> {messages,[{'$gen_cast',stop}]},
>>>            {links,[<0.3541.0>,<0.85.0>]},
>>>            {dictionary,[]},
>>>            {trap_exit,false},
>>>            {status,runnable},
>>>            {heap_size,233},
>>>            {stack_size,9},
>>>            {reductions,32}]}]]}}
>>>
>>> Any ideas?
>>>
>>> K.
>>> ---
>>> http://blitz.io
>>> http://twitter.com/pcapr
>>>
>>
>>
>>
>> --
>> Filipe David Manana,
>> fdmanana@gmail.com, fdmanana@apache.org
>>
>> "Reasonable men adapt themselves to the world.
>>  Unreasonable men adapt the world to themselves.
>>  That's why all progress depends on unreasonable men."
>>
>



-- 
Filipe David Manana,
fdmanana@gmail.com, fdmanana@apache.org

"Reasonable men adapt themselves to the world.
 Unreasonable men adapt the world to themselves.
 That's why all progress depends on unreasonable men."

Mime
View raw message