[Zope] Zope crashing a RedHat installation?
Dieter Maurer
dieter at handshake.de
Sat Oct 22 15:37:50 EDT 2005
Aaron Bauman wrote at 2005-10-21 15:50 -0400:
>We're running Zope 2.7.3, Plone 1.0.something
>relatively large Data.fs ( > 300MB )
>Decent amount of anonymous traffic,
>But relatively low administrative use (1 - 2 users usually).
>
>Could these factors be brining the machine down?
Up to now, I only met two software factors bringing a machine down:
* an unsocial process eats all available memory
Usually, the unsocial process is killed but the problem
may hit essential services before the unsocial process.
In this case, your system may become unusable.
An easy protection is the limitation of the
amount of virtual memory available to the process.
* an unsocial process creates an unbound amount of subprocesses
at a fast rate
Zope usually does not do the second type of thing (although, you
could write such an application).
In the past, Zope occationally was plagued by memory leaks -- usually
of the benign form (slowly leaking memory).
If you use a Zope application with a leak and you do not limit or supervise
memory consumption, then Zope might bring down your site.
But the probability is low...
>I've been getting intermittent 'corrupted data' messages for some time, but
>only recently been experiencing serious downtime.
Who reported '"corrupted data" messages'? What data is affected?
Usually, such messages (if issued by the OS) suggest a hardware
(and not a Zope) problem. Hardware problems may cause serious downtime, too.
>Is it time to seriously look at cleaning up the ZODB and upgrading?
The ZODB has almost surely nothing to do with this (unless
the '"corrupted data" messages' tell you that the ZODB storage is corrupt;
even then, hardware problems are more likely that software one).
>Or is my web host just trying to sell me more hardware?
You report *FAR* too few details about your problem as that
you could expect an answer to this question.
--
Dieter
More information about the Zope
mailing list