Dieter Maurer wrote:
This does not look like a crash.
The message tells you that a start failed because another running (!) process still has the ZODB storage file locked.
Almost surely, Zope is not dead (at least not completely).
I have seen similar behaviour in the past: caused by a Python threading bug triggered by a Linux 2.4 signal handling peculiarity:
A fatal signal kills the main Zope thread (handling all IO with the clients) but let the remaining threads in an insane state (could only be removed with "kill -9").
This is exactly what happens! The process still exists but it doesn't respond.
Current Python versions should have fixed this problem.
Operating system is SuSE Linux 8.0, Zope version is 2.5.0, Python 2.2,
Almost surely, one of the configurations with the bug mentioned above...
Well, for testing I tried to copy the Zope DB (Data.fs.*) into a fresh OpenSuSE 10.0 installation. Zope version there is 2.7.7. It works but now I get an authorization request dialog on every page. Any idea what could have gone wrong during the upgrade? -- Sincerely yours Sven Jacobs