Those Eternal Zope Errors that noone seem to know how to avoid - Why do they happen?
Hello! Can someone explain this to me: Zope 2.4.1 SunOS 5.8 Generic_108528-05 sun4u sparc Python 2.1.1 --without-pymalloc DCO2 latest from release TransparentFolders 0.3.2 Oracle 8.1.7 We keep getting these errors: ------ 2001-11-05T21:07:10 INFO(0) Z2 CONFLICT Competing writes at, /ka/acl_users/manage_edit Traceback (innermost last): File /usr/local/zope/dist/Zope-2.4.1/lib/python/ZPublisher/Publish.py, line 175, in publish File /usr/local/zope/dist/Zope-2.4.1/lib/python/Zope/__init__.py, line 240, in commit File /usr/local/zope/dist/Zope-2.4.1/lib/python/ZODB/Transaction.py, line 302, in commit File /usr/local/zope/dist/Zope-2.4.1/lib/python/ZODB/Connection.py, line 324, in commit ConflictError: '\x00\x00\x00\x00\x00\x006\xc7' ------ 2001-11-05T18:57:51 PROBLEM(100) ZServer unhandled connect event + a bunch of asyncore.py related errors The question - are we the only one experiencingthese errors or is anyone else having them? And what to do to avoid them? /dario - -------------------------------------------------------------------- Dario Lopez-Kästen Systems Developer Chalmers Univ. of Technology dario@ita.chalmers.se ICQ will yield no hits IT Systems & Services
There are others. :-) Search the mailing list archives for a) competing writes b) conflict errors Interesting that it's "acl_users/manage_edit" that's causing them. This might be something for the Collector... Danny On Tuesday 06 November 2001 10:40, Dario Lopez-Kästen wrote:
Hello!
Can someone explain this to me:
Zope 2.4.1 SunOS 5.8 Generic_108528-05 sun4u sparc Python 2.1.1 --without-pymalloc DCO2 latest from release TransparentFolders 0.3.2 Oracle 8.1.7
We keep getting these errors:
------ 2001-11-05T21:07:10 INFO(0) Z2 CONFLICT Competing writes at, /ka/acl_users/manage_edit Traceback (innermost last): File /usr/local/zope/dist/Zope-2.4.1/lib/python/ZPublisher/Publish.py, line 175, in publish File /usr/local/zope/dist/Zope-2.4.1/lib/python/Zope/__init__.py, line 240, in commit File /usr/local/zope/dist/Zope-2.4.1/lib/python/ZODB/Transaction.py, line 302, in commit File /usr/local/zope/dist/Zope-2.4.1/lib/python/ZODB/Connection.py, line 324, in commit ConflictError: '\x00\x00\x00\x00\x00\x006\xc7'
------ 2001-11-05T18:57:51 PROBLEM(100) ZServer unhandled connect event
+ a bunch of asyncore.py related errors
The question - are we the only one experiencingthese errors or is anyone else having them? And what to do to avoid them?
/dario
- -------------------------------------------------------------------- Dario Lopez-Kästen Systems Developer Chalmers Univ. of Technology dario@ita.chalmers.se ICQ will yield no hits IT Systems & Services
_______________________________________________ Zope maillist - Zope@zope.org http://lists.zope.org/mailman/listinfo/zope ** No cross posts or HTML encoding! ** (Related lists - http://lists.zope.org/mailman/listinfo/zope-announce http://lists.zope.org/mailman/listinfo/zope-dev )
Interesting that it's "acl_users/manage_edit" that's causing them. This might be something for the Collector...
hm... yes, in this particular case it is a combination of DCO2 weirdness and XUF weirdness, though I'd lay most of the blame on DCO2... there is something weird going on with transcations and DCO2, but I don't have the skills enough to pinpoint the exakt problem and report it to MattK. This is not the only place we've had these problems... I know that Steve Alexander has done some work on improving the transaction mecahnism in Zope, but I don't know if it'll find it's way in the Zope core... Still, would e mice to know how they occur.... will search the archives as you suggested. Re: asyncore problems, these are old problems (did a searchon the web), no solution so far... /dario
Re: asyncore problems, these are old problems (did a searchon the web), no solution so far...
I'm not sure that these particular asyncore.py errors are the same ones we've had in the past, but when we had "issues" with asycore.py a few months ago, we were able to track it down to poor signal handling in that particular python module. Older versions of Zope used their own asyncore, which had a loop on select.select, catching EINTR (a perfectly normal signal raised by the operating system under normal conditions). You may notice in your asyncore, that we just call select.select (line 86). I understand that this is just poor planning by the authors. I'm using a "fixed" asyncore.py, (one that catches EINTR and retries the select, and haven't gotten any errors since I started using it. HTH Ziniti
participants (3)
-
Danny William Adair -
Dario Lopez-Kästen -
John Ziniti