protocol bugs in memcached-1.1.13-pre(2)?

[Serhat Sakarya]

I've been able to reproduce this on 1.1.13, but not on 1.2.0. It looks like
an obscure edge case.

Regards,

Serhat

On 8/21/06, Brad Fitzpatrick <brad at danga.com> wrote:
>
> How easily can you reproduce this?  If you have a test script to
> demonstrate it, I'd love to find which commit broke it for you.
>
> Also, does it happen on the facebook branch?  (the 1.2.0-rc1 version)
>
>
>
> On Fri, 14 Jul 2006, Serhat Sakarya wrote:
>
> > I'm trying to integrate memcached into our relatively busy site (6M
> > pageviews/day), so I've been stress testing it on our network a bit.
> >
> > I was trying the latest download (1.1.13-pre2) and found a protocol
> > violation: sometimes the final "END" is missing, which causes problems
> for
> > e.g. the PHP memcache module -- it will report the node as down. This
> occurs
> > mostly when working with large data sets (300k+) from a process that
> runs on
> > the same machine. If a remote machine is used, it works fine...
> >
> > This problem occurs in 1.1.13-pre, but not in 1.1.12.
> >
> > I guess I should stick to 1.1.12 for my site then...?
> >
> > Regards,
> >
> > Serhat
> >
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.danga.com/pipermail/memcached/attachments/20060917/bb8a6321/attachment.html