Hi. I'm stuck to find why my Zope server sometimes hangs. STUPID_LOG_FILE sais me nothing (just no any error), Zope processes/instances are sleeping, host pings fine, but if you try to connect to the Zope with a simple browser -- just timeout and "bye-bye". What I need to switch on to debug this annoyance? After Zope restart (in console) all works fine. -- Regards, Bogdan "[In 'Doctor' mode], I spent a good ten minutes telling Emacs what I thought of it. (The response was, 'Perhaps you could try to be less abusive.')" (By Matt Welsh)
Hmm... I found some really strangeness after Zope restart. It always appears me in stupid_log_file after manager is logged in into ZMI. What's this? Does anymobdy know why mouned DB at /temp_folder goes crazy?: ......... File /opt/zope/lib/python/ZODB/Mount.py, line 43, in parentClassFactory File /opt/zope/lib/python/ZODB/Mount.py, line 43, in parentClassFactory File /opt/zope/lib/python/ZODB/Mount.py, line 43, in parentClassFactory File /opt/zope/lib/python/ZODB/Mount.py, line 43, in parentClassFactory File /opt/zope/lib/python/ZODB/Mount.py, line 43, in parentClassFactory RuntimeError: maximum recursion depth exceeded ------ 2002-12-03T08:48:37 INFO(0) ZODB Closed database: '142073576_1034688042.325979' ------ 2002-12-03T08:48:46 INFO(0) ZODB Opening database for mounting: '142073576_1034688042.325979' ------ 2002-12-03T08:48:46 INFO(0) ZODB Mounted database '142073576_1034688042.325979' at /temp_folder -- Regards, Bogdan This is an object-oriented system. If we change anything, the users object.
Bo M. Maryniuck writes:
I'm stuck to find why my Zope server sometimes hangs. STUPID_LOG_FILE sais me nothing (just no any error), Zope processes/instances are sleeping, host pings fine, but if you try to connect to the Zope with a simple browser -- just timeout and "bye-bye". What I need to switch on to debug this annoyance? Search the mailing list archives for "Big-M logging", "request profiler" or the "-M" option.
Before your Zope hangs completely, you may see hanging requests via "Control_Panel --> Debug Information". At the bottom, you see the currently executing requests. Dieter
On Tuesday 03 December 2002 22:31, Dieter Maurer wrote:
Before your Zope hangs completely, [skipp] Errmm.... How I can log-in to *dead* Zope, then go to the Control_Panel and see a Debug Information there? ;-)
-- Regards, Bogdan Colorless green ideas sleep furiously.
Before your Zope hangs completely, [skipp] Errmm.... How I can log-in to *dead* Zope, then go to the Control_Panel a= nd=20 see a Debug Information there? ;-)
hm, good question ;-) maybe do not start Zope at all, but load it via the debugger; see e.g.: http://www.zope.org/Members/mcdonc/HowTos/UsingTheZopeDebugger http://www.zope.org/Members/michel/HowTos/TheDebuggerIsYourFriend (or something else You get by a search for 'debugger' on zope.org) I hope You can 'import Zope' from the debugger commandline without hang, and then investigate why the requests cause zope to hang by stepping through the equivalent python call. HTH, clemens
On Wednesday 04 December 2002 13:39, Clemens Robbenhaar wrote:
I hope You can 'import Zope' from the debugger commandline without hang, and then investigate why the requests cause zope to hang by stepping through the equivalent python call.
What is debuggin Zope -- I know. ;-) The problem is that debugging shows nothing for me -- all is fubar always I debug. The only thing I did month ago is mounted another one data.fs, where is also another one "Control_Panel", "temp_folder", "Transient Object Container" where session data is stored and a "Browser Id Manager". Why I did this? Because Zope uses RAM for sessions and after Zope restart session is gone. But when you point a session to the main Data.fs -- "Undo" feature is gone! But I need *both of them* for my developers. What I think is probably my mount way was not too correct -- it works, but probably not so good. Since I'm not so sure of the problem -- I ask Zope gurus. ;-) -- Regards, Bogdan BASIC: A programming language. Related to certain social diseases in that those who have it will not admit it in polite company.
participants (3)
-
Bo M. Maryniuck -
Clemens Robbenhaar -
Dieter Maurer