Red Hat nash problem
Bill
ngc4013 at cox.net
Fri Sep 10 19:19:08 UTC 2004
>
>Message: 1
>Date: Fri, 10 Sep 2004 12:00:08 +0100
>From: Paul Furness <paul.furness at vil.ite.mee.com>
>Subject: Re: Red Hat nash problem
>To: For users of Fedora Core releases <fedora-list at redhat.com>
>Message-ID: <1094814007.13782.27.camel at zebra.vil.ite.mee.com>
>Content-Type: text/plain
>
>Hmmm. It looks like it's taking its time detecting the SCSI devices. All
>AIC79xx controllers, interestingly enough, although some of the servers
>are Dell (poweredge 1650, 1750 and 2650s) and some are completely
>different manufacturers.
>
>Thanks for the tip - I hadn't really though about it and therefore
>didn't appreciate what "quiet" meant...
>
>
>
>On Wed, 2004-09-08 at 14:09, Alexander Dalloz wrote:
>
>
>>Am Mi, den 08.09.2004 schrieb Paul Furness um 10:49:
>>
>>
>>
>>>If it's any help, I've recently done a lot of installs of FC2 on our
>>>number crunching cluster and a few workstations, and I've noticed that
>>>on a dual CPU system, the boot up appears to hang at the "Red Hat
>>>nash..." line for a considerable period - over a minute on many of the
>>>systems - before boot up continues normally.
>>>
>>>Occasionally - about 1 in 6 new builds - the system does actually hang
>>>at that point (> 30 minutes left to itself), but on all occasions (so
>>>far) a power cycle gets past it.
>>>
>>>
>>>Paul.
>>>
>>>
>>It would be interesting to know what you see when booting with "quiet"
>>mode off. Means erasing the "quiet" parameter from the kernel boot line.
>>This sets the verbosity back to normal.
>>
>>Alexander
>>
>>
>>
>
>
>
I have had very similar problems with a dual Athlon MP servers on Tyan
S2468 and S2469 motherboards. I have come to the same conclusion that
the delays in nash are related to the U320 SCSI interfaces and have
noticed many SCSI reset messages in the log files after running dmesg.
However, power cycling did nothing for mine. Here is another
insteresting tidbit of info.... I have one drive with AS-3 installed on
it and it shows no problems except of course that it will not recognize
my reiserfs RIAD boxes, one drive with FC1 and it works fine except for
the nash delays relating to the aic79xx driver, and another drive with
FC2 which has a tremendous hang time >15min on nash.
Oh yea... here is the kicker... with SUSE-9.1 all hardware configuration
combinations boot fine without any hardware hang ups durring SCSI driver
setup. Sadly the people where I work will not allow SUSE, so even I
would like to see this aic79xx driver problem fixed for FC2.
Also the nash hang problem only seems to happen when there are any U320
SCSI interfaces involved such as the Tyan S2469 which has U320 and I
have tried Adaptec U320 add on cards too, and the aic79xx driver is
being loaded.
Bill Cronk
More information about the fedora-list
mailing list