Redhat 7.2 kernel-smp-2.4.9-7 failure on 2550 Steve_Boley at
Mon Nov 5 10:32:00 CST 2001

You will need to load the CLI and use a container command to set the cache
to write back.  Here's a tip - If you are running a mail server or a
database (anything that is database oriented) you really need seperate
containers for your operating system and where your database is going to
exist.  Have write back on your os container and write through on the
database container.  Databases when pounded hard tend to flood the cache of
the controller and lock up container and corrupt the data.  Seems also to be
an issue with the new journaling file systems which are database oriented as
well.  The command to set the cache is 
AFA0>container set cache /write_cache_enable=TRUE X

X being the container number you want to set the cache setting on.

Also use: 

AFA0>container show cache X
to see what the cache settings are set to.

AFA0>controller battery_recondition /always=TRUE
is command to recondition the battery.
-----Original Message-----
From: Woodcock, Steve [mailto:SWoodcock at]
Sent: Monday, November 05, 2001 10:05 AM
To: Matt_Domsch at
Cc: Woodcock, Steve; linux-poweredge at
Subject: Re: Redhat 7.2 kernel-smp-2.4.9-7 failure on 2550

Matt_Domsch at writes:

> > Our 2550 fails to boot with latest smp kernel errata
> > (kernel-smp-2.4.9-7). It says:
> Latest is actually 2.4.9-13 now.   Progress...

Yup, I'm trying that one now. (My original message is a few days old
because I sent it to the wrong address and only resent it this

> > PCI: Found IRQ 9 for device 02:04.0
> > IRQ routing conflict for 02:04.0, have irq 10, want irq 9
> Hmmm.  That's kind of wierd.  What BIOS are you running?  A05 for
> the latest for the PE2550.

BIOS is A05

> > AAC:Batte
> This indicates that you need to perform a battery recondition (and
> that the driver has a bug in the printing of that message, but
> that's now understood).

Ok. I notice in the <ctrl-A> BIOS for the 3/DI that the write cache is
disabled, and the battery condition was "fine" or words to that
effect. I seem to remember some discussion here about this, can't
remember what the upshot was: can we enable the write cache? (Sorry,
this is a separate question really)

I'll try and get a dmesg from the failing smp kernel - I take it I
need to set the serial boot option in the BIOS and attach something
like HyperTerminal (Windows) to it? (We haven't got any proper serial

Regards, Steve Woodcock
Linux-PowerEdge mailing list
Linux-PowerEdge at

More information about the Linux-PowerEdge mailing list