On Fri, 8 Dec 2000, Luke Tymowski wrote:
Hello,
I've just tried and failed to upgrade to 2.2.4 from 2.2.2. A python instance starts up and slowly swallows the CPU, ending with 99% of the CPU. A second python instance starts and doesn't do anything naughty.
If I issue a ./stop, the second python instance ends properly, but the first continues to consume all CPU resources.
I've tried recompiling it several times, reinstalling Python, but no go. It did once start up properly, but as soon as I tried logging into Zope, python took over the CPU.
I'm using OpenBSD 2.8, python 1.5.2, x86, 196MB of RAM, GBs of free disk space.
I've been using Zope since 2.0 and have never had this much trouble upgrading. the only Zope product I'm using is SiteAccess (the same setup as the old Zope.org before the cluster upgrade).
Any ideas?
Try turning on the profiler. Quothe the control panel: To enable profiling, restart the Zope process with the environment variable PROFILE_PUBLISHER defined. The value of this variable should be the full system path to a file that will be used to dump a profile report when the process restarts or exits. And kill the process. Maybe this will create a profile report that tells you where it's spinning. Are you sure your waiting long enough? If you have a large Data.fs and you just move it over to a new install, Zope needs to take some time to build the initial database index file. -Michel