zope (with cmf) instance being frequently restarted by signal SIGSEGV(11)
Hi, I have a cmf site that has begun restarting frequently and at random intervals for no apparent reason. This first occurred about 2 weeks ago for about 2-3 days with the instance being restarted every 15 mins at worst. This problem disappeared as quickly as it had appeared. 2-3 days ago this same problem reappeared but with the site sometimes being restarted 10 times in a 15 minute period. I can find no sequence of url calls or any common reason for this to be occurring. The only thing I could find was the following entry appearing in the event log matching the times I have witnessed the instance restarting via top. ERROR(200) zdaemon Process 25606 terminated by signal SIGSEGV(11) The only changes that have occurred around the time of this first problem turning up was (other than normal site operation with members/items being add/deleted): -the cpu and memory being upgraded. This was returned to the original memory and cpu state to make sure this wasn't the cause. These have been swapped in/out with no change so it looks like it is just coincidence. -some site zope images were modified to set the 'precondition' attr to call a python script to set response headers for caching of images in apache... This hasn't been undone as the data.fs has since been packed. Is there any way this could be related??? There is no obvious signs of anything malicious going on, but this is always a possibility I guess? What to check for? As this has basically brought the site to its knees with proxy errors displayed more often than not, has anyone got suggestions on what to look for. possibilities (however remote) as I am pretty desperate to get this fixed. If there is a simple explanation, what could cause it to start out of the blue? The only possible answer I have come up with while searching is about the thread stack size being to small. And having to recompile python with this set to a larger figure. Is this related to the problem I am having, and is there an easy way to check if this is the case for the version I have running? If this is what I need to do, is there a good link for details on what needs to be done and how to do it? CMF: 1.4.2 Zope: (unreleased version, python 2.1.3, linux2) Python: 2.1.3 (#1, Sep 7 2002, 15:29:56) [GCC 2.95.4 20011002 (Debian prerelease)] Thank you for all your help on this as I am lost at what steps to take next. Thanks, Richard.
Richard Ettema wrote:
2-3 days ago this same problem reappeared but with the site sometimes being restarted 10 times in a 15 minute period.
Well, looks like you need to find out what's changed in the times where this happens lots, your serverand Apache logs may help here...
matching the times I have witnessed the instance restarting via top.
ERROR(200) zdaemon Process 25606 terminated by signal SIGSEGV(11)
Yeah, that tells you something died. Killed by a SIGSEGV(11). What Zope version are you running? What Python version are you running? What OS are you running on? What versions of what products are you running? Do any of them use C extensions to python?
-some site zope images were modified to set the 'precondition' attr to call a python script to set response headers for caching of images in apache... This hasn't been undone as the data.fs has since been packed. Is there any way this could be related???
Maybe, try taking this out of the equation...
There is no obvious signs of anything malicious going on, but this is always a possibility I guess? What to check for?
Apache logs for odd URLs...
The only possible answer I have come up with while searching is about the thread stack size being to small. And having to recompile python with this set to a larger figure. Is this related to the problem I am having, and is there an easy way to check if this is the case for the version I have running?
What OS are you running on?
CMF: 1.4.2 Zope: (unreleased version, python 2.1.3, linux2)
Ack. That might be Zop 2.7.0, is it? In which case...
Python: 2.1.3 (#1, Sep 7 2002, 15:29:56) [GCC 2.95.4 20011002 (Debian prerelease)] ...why aren't you running on Python 2.3?
cheers, Chris -- Simplistix - Content Management, Zope & Python Consulting - http://www.simplistix.co.uk
participants (2)
-
Chris Withers -
Richard Ettema