couchdb-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Jason Smith <...@iriscouch.com>
Subject Re: Occasional corrupt request
Date Wed, 11 Jan 2012 08:55:46 GMT
Hi, Pete. Does this error happen only (or mostly) when your request to
couch contains a body? Is it impossible to reproduce using CouchDB
1.0? Does it happen most often when attempting a privileged operation
which you are not permitted to do? Are you able to run a packet
capture like tcpdump and confirm if couch is sending any RST packets
to the client?

If the answer to all of these is "yes" then I have seen this too. I
struggled with it but finally used a workaround. I would welcome any
additional clues.

If the answer is "no" then, sorry, I myself haven't seen that yet. My
standard tool for this type of problem is tcpdump, FWIW.

On Wed, Jan 11, 2012 at 1:14 PM, Pete Vander Giessen <petevg@gmail.com> wrote:
> Hi All,
>
> We've had a sporadic problem, and I haven't been able to put together
> a consistent repro case, so I'm hesitant to file a bug. But I did come
> across some old bugs that seemed relevant, and had a couple of
> questions.
>
> The problem: we're using couchdbkit to talk to a local couch server.
> Every once in a while, a request fails, and it looks like it fails
> because we're getting data for a different request in the socket. We
> submitted a fix for a potential race condition in couchdbkit's socket
> pooling last year, and that made the problem much less frequent, but
> we still see it occasionally.
>
> My question: COUCHDB-95 is an old bug, but it seems to describe
> something like what we're seeing. SVN r679636 fixed the issue, but I'm
> having a little trouble understanding exactly what the fix did. Is it
> a workaround for a bug in mochiweb, or a bug in couch itself? Can
> anybody walk me through the fix, and let me know whether it might be
> relevant to couchdbkit?
>
> Also: is anybody else seeing similar issues? When it occurs,
> couchdbkit throws a NoMoreData, InvalidChunkSize, or ParserError
> exception (this last when it doesn't realize that the data it has is
> corrupted until it closes the connection and tries to parse the JSON).
> I've got stopgap code in place to retry a request on these exceptions,
> but I'd like to get to the heart of the issue.
>
> Thank you,
>
> ~PeteVG
>
> "The problem with Internet quotations is that many are not genuine."
> ~ Abraham Lincoln



-- 
Iris Couch

Mime
View raw message