On Wed, Aug 27, 2008 at 7:32 AM, Ingvar Hagelund <[EMAIL PROTECTED]> wrote:
> Now, writing to the ZEO database is very slow
Hard to tell the cause of that from here. Strace, tcpdump, and
wireshark are my friends.
> and the same goes for
> searching, which is a problem for the users that put information into
> the solution.
Larger ZODB object caches and ZEO disk caches might help there.
> Average cpu usage on the box running ZEO is some 10%, and never above
> 65% (of one cpu core).
Watching the output "iostat -x 60" (especially the % utilization) can be
helpful in identifying disk bottlenecks (which I'm not especially
suspicious of in this case).
> The size of Data.fs is quite large, about 8.2GB after packing.
I wouldn't worry about a Data.fs of that size. They can comfortably
grow much larger than that.
> Now for some questions:
> - Is a configuration like the one described a sensible setup?
> - Are there any ovious key tuning buttons that we have missed?
See the message from Marius for the most likely culprit.
> - Would it be an idea to move binaries out of Data.fs and serve them
> from a shared filesystem instead?
Maybe, maybe not; the details are important there. My suspicion is that
you wouldn't see much benefit.
Senior Software Engineer
For more information about ZODB, see the ZODB Wiki:
ZODB-Dev mailing list - ZODB-Dev@zope.org