couchdb-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Filipe David Manana <fdman...@apache.org>
Subject Re: HTTP Error 303 causes erlang stack trace
Date Wed, 13 Oct 2010 23:20:10 GMT
On Wed, Oct 13, 2010 at 11:00 PM, John Logsdon
<john.logsdon@netdev.co.uk> wrote:
> Hi David
>
> Many thanks for patching this.
>
> TBH I have no idea where the 303 came from so I will have to simulate it to
> prove the patch.

So it's something not easily reproducible? (I never got 303 when
replicating to/from couchone.com)

>
> My question though is:
>
> In an unexpected situation e.g. Unknown Error code. What should CouchDB do?
> I kind of expected that it would see that as a transient problem and try
> again later.

That's exactly what it normally does (for error codes 50x except 501).
This case was exceptional. The replicator also follows redirects (301
and 302 HTTP response codes).

>
> The reason I ask is that I love this technology but I'm selling it into a
> bunch of guys who have breathed relational DB's for 20 yrs and I just want
> it to be successful 'cause I know its the right solution for the problem
> we're trying to solve.
>
> Unfortunately new technologies get less slack than old ones with this kind
> of group!
>
> Thanks for your speedy reply.

You're welcome.
I'll wait a bit to see if you're able to reproduce or someone else.
Nevertheless, it's something definitely worth committing.

>
> Regards
>
> John
>
> On 13 October 2010 21:55, Filipe David Manana <fdmanana@apache.org> wrote:
>
>> John,
>>
>> I created a ticket, https://issues.apache.org/jira/browse/COUCHDB-918
>> and attached there a patch.
>>
>> Can you test the patch? It was made against trunk, but should apply
>> cleanly against branch 1.0.x as well.
>> I'll commit it afterwards.
>>
>> On Wed, Oct 13, 2010 at 8:57 PM, Filipe David Manana
>> <fdmanana@apache.org> wrote:
>> > Hi John,
>> >
>> > Do you have some proxy in the middle that might be returning the 303
>> > response? Perhaps in the meanwhile you can reconfigure your proxy.
>> > CouchDB itself doesn't send 303 errors anywhere. The crash is because
>> > it's not dealing with 303 (only 301 and 302).
>> >
>> > In the meanwhile, do you want to open a Jira ticket?
>> > I'll gladly do a patch and commit it.
>> >
>> > On Wed, Oct 13, 2010 at 2:42 PM, John Logsdon <johnlogsdon@me.com>
>> wrote:
>> >> Hello again
>> >>
>> >> Can anyone comment on this replication error I posted a few days ago? If
>> it appears to be a genuine bug then I can raise it through the project but I
>> just wanted some feedback before I did that.
>> >>
>> >> Regards
>> >>
>> >> John
>> >>
>> >> On 11 Oct 2010, at 17:47, John Logsdon wrote:
>> >>
>> >>> Hi
>> >>>
>> >>> I kicked off a replication to couchone.com from my local couchdb by
>> pushing from the local db.
>> >>>
>> >>> It was for a 2M doc database and all was going well up until 678K
>> entries but then I had a connection issue and well it's all gone horribly
>> wrong.
>> >>>
>> >>> Anyway basically the replication has gone from "Replicator" and I cant
>> restart it. The erlang stack trace in couchdb.log looks like this:
>> >>>
>> >>> ** Reason for termination ==
>> >>> ** {http_request_failed,<<"unhandled response code 303">>}
>> >>>
>> >>>
>> >>> [Sun, 10 Oct 2010 17:29:52 GMT] [error] [<0.17737.2>]
>> {error_report,<0.30.0>,
>> >>>     {<0.17737.2>,crash_report,
>> >>>      [[{initial_call,{couch_rep_reader,init,['Argument__1']}},
>> >>>        {pid,<0.17737.2>},
>> >>>        {registered_name,[]},
>> >>>        {error_info,
>> >>>            {exit,
>> >>>                {http_request_failed,<<"unhandled response
code 303">>},
>> >>>
>>  [{gen_server,terminate,6},{proc_lib,init_p_do_apply,3}]}},
>> >>>        {ancestors,
>> >>>
>>  [<0.17728.2>,couch_rep_sup,couch_primary_services,couch_server_sup,
>> >>>             <0.31.0>]},
>> >>>        {messages,
>> >>>            [{'EXIT',<0.17738.2>,
>> >>>                 {{http_request_failed,<<"unhandled response
code
>> 303">>},
>> >>>                  {gen_server,call,
>> >>>                      [<0.17735.2>,next_missing_revs,infinity]}}}]},
>> >>>        {links,[]},
>> >>>        {dictionary,[]},
>> >>>        {trap_exit,true},
>> >>>        {status,running},
>> >>>
>> >>>
>> >>> Have I bumped into a bug here or....????
>> >>>
>> >>>
>> >>> Regards
>> >>>
>> >>> John
>> >>
>> >>
>> >
>> >
>> >
>> > --
>> > 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."
>>
>
>
>
> --
> John Logsdon
> CEO
> NetDev Limited
> +44 1273 773661
> +44 7899 811245
> http://www.netdev.co.uk
> Registered in England and Wales
> Company Number 04741258
>



-- 
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