Workaroud for a SATA nForce 250 SCSI Parity Error problem
Hi everyone, I have the following configuration: MSI K8N NEO2-FX,nForce3 250Gb,Socket-939 Raid, ATX, SATA, GbLAN, Sound, AGP8X AMD Athlon 64 3000+ 1.8GHz Socket 939 1Gb kit DDR PC3200 Premier Dual Channel - OCZ XFX GeForce FX5500 128MB DDR, AGP8X, Tv-Out, DVI-I, Retail Maxtor DiamondMax 10 160GB SATA2 8MB 7200RPM LG DVD Drive, GDR-8164RBB, 16x52, Black OEM Eventually, not without struggling, I managed to install and run SLES 9 64bit + SP3 on this box (I use it as a very small office server running TFTP, FTP, Apache and Samba). The trick was to lower the CAS settings on the memory and use ide=nodma. So with these settings it was running fine for a month or so. Now I want to “upgrade” it to SUSE 10.0 or SUSE 10.1 64bit since I need some packages from those releases that seem to be a hassle to install them in SLES 9 (i.e. MIT Kerberos 5). The problem is that I just can’t install them. Tried with both and hit the same issue: once I select “update” option and the installer tries to mount existing Linux partitions, I can see on the logger that infamous “SCSI Parity Error”, “device busy”, timeout and I/O errors while accessing the hard drive. And the only way out is to use the POWER RESET button. SLES 9 seems to mount the SATA as an IDE (/dev/hdg1 thus I can use ide=nodma to get around the issue) but SUSE 10.X sees it as a SCSI drive (dev/sda) and thus ide=nodma. Both use sata_nv module. Now all I want is somehow install SUSE 10.X on it – don’t care about performance. Is it any way I can “force” SUSE10.X installer to “mount” the SATA disk as an IDE disk and not as a SCSI – the same way SLES 9 does ? Is there any updated sata_nv module I could use ? Or can I use any other “generic” module for this SATA chipset and drive ? Your thoughts on this matter will be very much appreciated. Thank you. Best regards, Florin
SuSE 10.1 uses a new enough kernel that you should be able to get help on the nodma question on *linux-ide@vger.kernel.org* (The official Linux Kernel SATA devel and support list). * *I don't thnk you have to subscribe to post and they follow a reply-all policy on all the linux kernel lists, so again no need to subscribe. Greg On 5/26/06, hatespam00-mlist@yahoo.com <hatespam00-mlist@yahoo.com> wrote:
Hi everyone,
I have the following configuration:
MSI K8N NEO2-FX,nForce3 250Gb,Socket-939 Raid, ATX, SATA, GbLAN, Sound, AGP8X AMD Athlon 64 3000+ 1.8GHz Socket 939 1Gb kit DDR PC3200 Premier Dual Channel - OCZ XFX GeForce FX5500 128MB DDR, AGP8X, Tv-Out, DVI-I, Retail Maxtor DiamondMax 10 160GB SATA2 8MB 7200RPM LG DVD Drive, GDR-8164RBB, 16x52, Black OEM
Eventually, not without struggling, I managed to install and run SLES 9 64bit + SP3 on this box (I use it as a very small office server running TFTP, FTP, Apache and Samba). The trick was to lower the CAS settings on the memory and use ide=nodma. So with these settings it was running fine for a month or so.
Now I want to "upgrade" it to SUSE 10.0 or SUSE 10.1 64bit since I need some packages from those releases that seem to be a hassle to install them in SLES 9 (i.e. MIT Kerberos 5).
The problem is that I just can't install them. Tried with both and hit the same issue: once I select "update" option and the installer tries to mount existing Linux partitions, I can see on the logger that infamous "SCSI Parity Error", "device busy", timeout and I/O errors while accessing the hard drive. And the only way out is to use the POWER RESET button.
SLES 9 seems to mount the SATA as an IDE (/dev/hdg1 thus I can use ide=nodma to get around the issue) but SUSE 10.X sees it as a SCSI drive (dev/sda) and thus ide=nodma. Both use sata_nv module.
Now all I want is somehow install SUSE 10.X on it – don't care about performance. Is it any way I can "force" SUSE10.X installer to "mount" the SATA disk as an IDE disk and not as a SCSI – the same way SLES 9 does ? Is there any updated sata_nv module I could use ? Or can I use any other "generic" module for this SATA chipset and drive ?
Your thoughts on this matter will be very much appreciated. Thank you.
Best regards, Florin
-- Check the headers for your unsubscription address For additional commands send e-mail to suse-linux-e-help@suse.com Also check the archives at http://lists.suse.com Please read the FAQs: suse-linux-e-faq@suse.com
-- Greg Freemyer The Norcross Group Forensics for the 21st Century
On Friday 26 May 2006 13:56, hatespam00-mlist@yahoo.com wrote:
mount existing Linux partitions, I can see on the logger that infamous “SCSI Parity Error”, “device busy”, timeout and I/O errors while accessing the hard drive. And the only way out is to use the POWER RESET button.
Where do I found more info on these infamous errors? I think the same bug has bit me. nforce 350 + maxtor drives = > drive failures, lots of bad sectors etc. Should they work on a different chipset than nvidia? One of those failed drives has very precious data on it. It's part on a RAID1 mirror. The other disk died completely.
On 5/26/06, Silviu Marin-Caea <silviu_marin-caea@fieldinsights.ro> wrote:
On Friday 26 May 2006 13:56, hatespam00-mlist@yahoo.com wrote:
mount existing Linux partitions, I can see on the logger that infamous "SCSI Parity Error", "device busy", timeout and I/O errors while accessing the hard drive. And the only way out is to use the POWER RESET button.
Where do I found more info on these infamous errors?
I think the same bug has bit me.
nforce 350 + maxtor drives = > drive failures, lots of bad sectors etc.
Should they work on a different chipset than nvidia?
One of those failed drives has very precious data on it. It's part on a RAID1 mirror. The other disk died completely.
The latest -mm vanilla kernel has a huge SATA driver rework in it. The goal is to get that rework into 2.6.18 in a few months, and then I hope into SUSE 10.2. Until then you could review the linux-ide mailing list archives and see if it looks like the bif rework will help you or not. If so, the big rework patch has also been backported to one or two stable (vanilla) kernels. You could consider using a stable vanilla kernel with the rework patch and see if it helps. Greg -- Greg Freemyer The Norcross Group Forensics for the 21st Century
Thanks very much Greg, Regards, Florin ----- Original Message ---- From: Greg Freemyer <greg.freemyer@gmail.com> To: Silviu Marin-Caea <silviu_marin-caea@fieldinsights.ro> Cc: suse-linux-e@suse.com Sent: Friday, May 26, 2006 1:48:23 PM Subject: Re: [SLE] Workaroud for a SATA nForce 250 SCSI Parity Error problem On 5/26/06, Silviu Marin-Caea <silviu_marin-caea@fieldinsights.ro> wrote:
On Friday 26 May 2006 13:56, hatespam00-mlist@yahoo.com wrote:
mount existing Linux partitions, I can see on the logger that infamous "SCSI Parity Error", "device busy", timeout and I/O errors while accessing the hard drive. And the only way out is to use the POWER RESET button.
Where do I found more info on these infamous errors?
I think the same bug has bit me.
nforce 350 + maxtor drives = > drive failures, lots of bad sectors etc.
Should they work on a different chipset than nvidia?
One of those failed drives has very precious data on it. It's part on a RAID1 mirror. The other disk died completely.
The latest -mm vanilla kernel has a huge SATA driver rework in it. The goal is to get that rework into 2.6.18 in a few months, and then I hope into SUSE 10.2. Until then you could review the linux-ide mailing list archives and see if it looks like the bif rework will help you or not. If so, the big rework patch has also been backported to one or two stable (vanilla) kernels. You could consider using a stable vanilla kernel with the rework patch and see if it helps. Greg -- Greg Freemyer The Norcross Group Forensics for the 21st Century
Hi Silviu, You'll find quite a few if you google for "sata_nv" and "scsi parity error" There are problems with VIA chipsets as well but at least on my home computer (with VIA) the sata drive is loaded as an IDE and I can use ide=nodma to get around the issue - the HDD access won't be very fast but at least it works without lockups. This is not the case however with nForce 250 chipset where ide=nodma is "invalid" and thus ignored. Haven't tried with other nVidia chipsets though. Strangely enough that SLES 9 will work OK as long as DMA is disabled ... Regards (aka numa cele bune :-) Florin ----- Original Message ---- From: Silviu Marin-Caea <silviu_marin-caea@fieldinsights.ro> To: suse-linux-e@suse.com Sent: Friday, May 26, 2006 1:35:41 PM Subject: Re: [SLE] Workaroud for a SATA nForce 250 SCSI Parity Error problem On Friday 26 May 2006 13:56, hatespam00-mlist@yahoo.com wrote:
mount existing Linux partitions, I can see on the logger that infamous “SCSI Parity Error”, “device busy”, timeout and I/O errors while accessing the hard drive. And the only way out is to use the POWER RESET button.
Where do I found more info on these infamous errors? I think the same bug has bit me. nforce 350 + maxtor drives = > drive failures, lots of bad sectors etc. Should they work on a different chipset than nvidia? One of those failed drives has very precious data on it. It's part on a RAID1 mirror. The other disk died completely. -- Check the headers for your unsubscription address For additional commands send e-mail to suse-linux-e-help@suse.com Also check the archives at http://lists.suse.com Please read the FAQs: suse-linux-e-faq@suse.com
participants (4)
-
Florin -
Greg Freemyer -
hatespam00-mlist＠yahoo.com
-
Silviu Marin-Caea