> > I have a Zope 2.5.1 using ZEO on Linux, with Apache as proxy server.
> > ZEO clients run fine for a couple of hours, and then suddenly one of
> > stops responding. It's the type of Zope hangs when all processes are
> > but ZServer just doesn't respond (load is very low, with the CPU 90%
> > so it is not a performance issue or recursion problem).
> > There is no error before this situation happens. The processes just
> > respond any more. The second ZEO keeps running, but that machine also
> > from time to time, so I guess ZEO is not the problem, and the hardware
> > any Linux-specific problems aren't either.
> > When the ZServer hangs, I can not access it directly, so it is no
> > with the Apache.
> > You see, I don't have any clue on what is happening here. The servers
> > fine for months, and I don't know what specific change could have
> > problem.
> > How can such a situation be debugged?
Dieter Maurer wrote:
> Maybe, you use the "-M" logging and analyse the log with Chris'
> "requestprofiler". You may see unfinished requests.
Thanks! That's what I did. I think I found the problem: I have some Objects
that inherit from Image, and their "data" attribute (that's where the image
data is held) wasn't initialized correctly in some cases. So whenever
somebody tried to display one of those objects, Zope started streaming back
the image data, which made the request hang when the data attribute was
Zope-Dev maillist - [EMAIL PROTECTED]
** No cross posts or HTML encoding! **
(Related lists -