|
Hi all,
I'm not 100% sure this is relevant, but we had a problem with the 2.6
kernel and wait/signal handlers
in some other project. I know RHEL3 has some 2.6 features backported into
it's 2.4 kernel, do this
may be it.
It seems the order in which processes can run after a fork has changed in
2.6. From what I understand,
after a fork, the child now (can?) runs first instead of the parent (or
something like that). This
means that it's now possible for the child to terminate even before the
parent was a chance to wait for it. Then if your handler is IGNORE you
will never get a chance to reap the child.
Does this make any sense in this context?
Patrick
----- Original Message -----
> On Mon, Nov 08, 2004 at 10:22:18AM +0000, Johan@ Elmerfjord wrote:
>> Hi Horms,
>>
>> Setting 'local $SIG{'CHLD'} = undef;' works just as good as
>> outcommenting the line. So I will go for this solution.
>> It seems more foolprof to have it set to something.. So it's not
>> completly up to the version of perl currently intalled.. :-)
>
> I think that if you uncomment it, then the signal handler may
> potentially come into play from time to time, leading to weird/bad
> things. So the undef solution is good IMHO. I will put it into CVS
> in the morning.
>
> --
> Horms
> _______________________________________________
> LinuxVirtualServer.org mailing list - lvs-users@xxxxxxxxxxxxxxxxxxxxxx
> Send requests to lvs-users-request@xxxxxxxxxxxxxxxxxxxxxx
> or go to http://www.in-addr.de/mailman/listinfo/lvs-users
>
>
|
