Proliant 5000 sever Fbsd 5.4 (re)boot problem

[Robert Slade]

On Wed, 2005-08-31 at 00:21, Vizion wrote:
> On Tuesday 30 August 2005 13:31,  the author Robert Slade contributed to the 
> dialogue on-
>  Re: Proliant 5000 sever Fbsd 5.4 (re)boot problem: 
> 
> >On Tue, 2005-08-30 at 21:20, Vizion wrote:
> >> On Tuesday 30 August 2005 13:22,  the author Robert Slade contributed to
> >> the dialogue on-
> >>
> >>  Re: Proliant 5000 sever Fbsd 5.4 (re)boot problem:
> >> >On Tue, 2005-08-30 at 20:10, Vizion wrote:
> >> >> On Tuesday 30 August 2005 12:05,  the author Robert Slade contributed
> >> >> to the dialogue on-
> >> >>
> >> >>  Proliant 5000 sever Fbsd 5.4 (re)boot problem:
> >> >> >Hiya,
> >> >> >
> >> >> >I've been working on this beasty on and off for some time. It's a Quad
> >> >> >processor 1 Gbyte of memory and 5 scsi drives using the 2p raid
> >> >> >controller setup as 2 raid arrays + 1 spare.
> >> >> >
> >> >> >The machine works fine with 5.4 release #0 with the supplied generic
> >> >> >kernel.
> >> >> >
> >> >> >The problem(s) I have been having are:
> >> >> >
> >> >> >1. When I recompiled the Kernel with SMP support, I get random
> >> >> > reboots. It also fails to boot sometimes failing at the point after
> >> >> > waiting for the scsi drives to settle. I get some error codes and
> >> >> > Fbsd fails to find the boot device.
> >> >> >
> >> >> >2. I CVSuped to 5.4 release #2 and recompiled the Kernel with SMP
> >> >> >support. This does to boot at all. It gets as far as the waiting 15s
> >> >> > for scsi devices to settle, then (appears to) reset the scsi
> >> >> > controller and immediately tries to access the drives (does not
> >> >> > wait). I have tried recompiling with scsi_delay set to 30000 (30s)
> >> >> > with no change.
> >> >> >
> >> >> >I have checked dmesg and message logs but there is nothing related to
> >> >> >the problem(s) there.
> >> >> >
> >> >> >I have gone back to the 5.4 release #0 single processor kernel for now
> >> >> >which is a shame as the machine is slow without the multi processor
> >> >> >support.
> >> >> >
> >> >> >The only thing out of the ordinary I have noted is a tx underunn --
> >> >> >increasing threshold to 512 bytes  message which appears related to
> >> >> >running kde remotely via vncserver and tinync.
> >> >> >
> >> >> >Any ideas, I can send conf files etc if needed.
> >> >> >
> >> >> >Thanks
> >> >> >
> >> >> >Rob
> >> >>
> >> >> What are your bios setting?
> >> >> My guess is that you have not made the right setting using the siftware
> >> >> and configuration utilities
> >> >> david
> >> >
> >> >David,
> >> >
> >> >Thanks. The BIOS setting appear ok - OS type is set as UNIX (Small disk
> >> >geometry) and the machine passes all the diagnostics.
> >> >
> >> >Rob
> >>
> >> I cannot remember - but I have sneaking notion that you need to set it as
> >> linux
> >
> >Tried that too :-). I think that the problem is that with 5.4 release #2
> >it is trying to access that scsi drives immediately then inducing the
> >kernel panic for 15s.  Rather than inducing the panic 1st.
> >
> >Rob
> >
> 
> Did you follow my suggestion and search the HP resources with freebsd and your 
> model.  I have had the same problem myself I am pretty certain it was fixed 
> by changes using the Proliant Essrntial Foundation Pack.. but my memory may 
> not be accurate.
> david

David,

I did update the system and controller ROMs whne the machine was running
windows. I have been on the HP site and as far as I can tell I have the
latest.

Rob