Just occured to me if Im leaking REQUEST it would be nice to see exactly which requests im leaking. But is that really covered by the -M logger, if a request fails to be closed in there does that really translate to a leaked request? I had a look in the code and it wasn't obvious to me.... Cheers. -- Andy McKay.
It's not covered by the data output into the -M log. Objects can leak without requests hanging or taking a long time. But you might be able to turn on profiling (set the PROFILE_PUBLISHER env var to a filename and restart Zope) on a *test* box (do not do it in production, it drastically effects performance), and look at the profile output. If the memory loss is severe, it might help to investigate the methods that have the largest "ncalls" value (the most frequently called methods). That said, there's no globally easy way to diagnose a memory leak. It's a matter of exclusion. ----- Original Message ----- From: "Andy McKay" <andym@ActiveState.com> To: <zope-dev@zope.org> Cc: <chrismc@digicool.com> Sent: Wednesday, April 25, 2001 1:08 PM Subject: [Zope-dev] Memory leak (2)
Just occured to me if Im leaking REQUEST it would be nice to see exactly which requests im leaking. But is that really covered by the -M logger, if a request fails to be closed in there does that really translate to a leaked request? I had a look in the code and it wasn't obvious to me....
Cheers. -- Andy McKay.
_______________________________________________ Zope-Dev maillist - Zope-Dev@zope.org http://lists.zope.org/mailman/listinfo/zope-dev ** No cross posts or HTML encoding! ** (Related lists - http://lists.zope.org/mailman/listinfo/zope-announce http://lists.zope.org/mailman/listinfo/zope )
It's not covered by the data output into the -M log. Objects can leak without requests hanging or taking a long time. But you might be able to turn on profiling (set the PROFILE_PUBLISHER env var to a filename and restart Zope) on a *test* box (do not do it in production, it drastically effects performance), and look at the profile output. If the memory loss is severe, it might help to investigate the methods that have the largest "ncalls" value (the most frequently called methods). That said, there's no globally easy way to diagnose a memory leak. It's a matter of exclusion.
Well there goes that idea. Forgot about profiling, thanks. -- Andy McKay.
participants (2)
-
Andy McKay -
Chris McDonough