From: Matthias Eble (matthias.eble_at_mailing.kaufland-informationssysteme.com)
Date: Fri Oct 06 2006 - 10:59:06 CEST
Message-ID: <45261ADA.8060502@mailing.kaufland-informationssysteme.com> Date: Fri, 06 Oct 2006 10:59:06 +0200 From: Matthias Eble <matthias.eble@mailing.kaufland-informationssysteme.com> Subject: Re: [suse-sles-e] random reboots
Penhale, Jeff schrieb:
> If these tools cause problems, then report it to the relevant vendor. These
> people know the most about their systems, so their tools should always be
> one of the first options.
this should be the way to go. I started an itrc thread a couple of
months ago - no valuable response - only an I've got the same problem post
>
> Vendors will insist on you running their tools before any support issues are
> raised with them.
I only know that for hp hardware this doesn't need to fit since the iml
is accessible through ILO as well.
>
> I mean it stands to reason, why bypass IBM support tools when diagnosing a
> problem on an IBM box, or HP tools with a problem with HP hardware.
the problem is that imo (and I'm not alone) the hp-psp is (sorry) a
bunch of ..
installing the complete pack ends up in a system running multiple
strange processes doing (sometimes) fancy things like blocking a cpu
with 100% user util until restarting the hpasmd.
kernel modules are added also and the hpsmh daemon isn't acceptable in
production to me.
We run a stripped down hpasmd (only running hpasmd; no cma* procs)
in combination with hpacucli monitored by nagios.
Yet hpasmd sometimes (strangely on 8 special dl585) pulls 100%, but
we're quite happy with our setup though.
>>
>> (That's why I don't like to install _ANY_ vendor-specific
>> _support packs_
>> /DELL, HP, etc... - it is always better to avoid them if possible/.
>> It is not possible /not to install/ in some cases, but it is _ALWAYS_
>> stability risk.)
the problem with that is that (in case of hp) power supply/fan failures
will not be detectable without hpasm. Disk failures not without hpacucli.
matthias
---------------------------------------------------------------------
To unsubscribe, e-mail: suse-sles-e-unsubscribe@suse.com
For additional commands, e-mail: suse-sles-e-help@suse.com
This archive was generated by hypermail 2.1.7 : Fri Oct 06 2006 - 10:59:28 CEST