On 30 Oct 2007 at 18:18, Brian Kroth wrote:
> It also seems strange to me that only nagios and cactid would crap out
> on this. If I run the cacti poller just through php I get no errors.
> The server also hosts ntp, portage over NFS, syslog-ng, mysql, dovecot,
> sec, postfix, and a bunch of management scripts for other servers. It's
> a busy thing, so if bad ram were the case, I'd expect to see them in
> other places. Possibly just bad code on cactid/nagios' part? Should
> PAX complain then?
PaX only logs bad behaviour when it's about unintended code execution,
this bug here is a kernel-internal consistency check, it doesn't mean
that the triggering userland app has a bug, it may very well just manage
to trigger a kernel bug (which in turn could be in PaX for all i know).
it'd be nice if you could find a way to reproduce it, also if you could
test with 2.6.23-pax.
> Also including emerge --info for anyone who's interested. Should I be
> opening a bug for this?
>
> # pcregrep " page:" /var/log/kern.log | sed 's/.*\(page:\S\+\).*/\1/' |
> perl -e 'while (<>) { $data{$_}++ }; for (keys %data) { print $data{$_}
> . " " . $_ ; };'
>
> 1 page:c14320e0
what is more interesting is the other fields, especially flags.
--
[EMAIL PROTECTED] mailing list
