[ZODB-Dev] return of corruption errors

[Tim Peters]

[Gerry Kirk]
> I notice a few of these in /var/log/messages:
>
> application bug: python2.3(23085) has SIGCHLD set to SIG_IGN but calls
wait()

> Could that be causing corruption?

I doubt it.  Because usable signals and a usable wait() don't exist on
Windows, nothing in ZODB/ZEO relies on either.  I'm sure this complaint was
about a Unix-specific zdaemon-based "ctl" script of some sort, and those
never muck with database files.

> Googling, I came across this in the zope collector:
>
> http://zope.org/Collectors/Zope/1336.

Which appears to be about the zeoctl script.

> I'm using zope 2.7.0, so I likely have that bug.

Sorry, I don't know -- but strongly doubt it's relevant to file corruption
problems.