On Thu, Mar 09, 2006 at 09:40:55AM -0500, Derrick Lobo wrote:
> I get these errors on my 300GB drives until the server crashes. The server
> is a HP Proliant DL385 with Netbsd 2.0(amd64) and an adaptec SCSI
> controller. Since its an amd64 server Im using this mail list...
> 
> Copied 8 bytes of sense data offset 12: 0xf0 0x0 0x1 0x13 0xbb 0x7b 0x5d
> 0x28
> sd1(ahd0:0:1:0):  Check Condition on CDB: 0x28 00 13 bb 7b 3f 00 00 80 00
>     SENSE KEY:  Recovered Error
>    INFO FIELD:  331053917
>      ASC/ASCQ:  No Additional Sense Information

Isn't the last error before crash different ? What is the panic message ?

> 
> Details about the SCSI adapter
> ahd0 at pci5 dev 7 function 0
> ahd0: Enabling 39Bit Addressing
> ahd0: interrupting at ioapic3 pin 2 (irq 10)
> ahd0: aic7901: Ultra320 Wide Channel A, SCSI Id=7, PCI-X 67-100Mhz, 512 SCBs
> scsibus0 at ahd0: 16 targets, 8 luns per target
> 
> Details about the drive
> sd1 at scsibus0 target 1 lun 0: <COMPAQ, BD3008856C, HPB2> disk fixed
> sd1: 279 GB, 78753 cyl, 8 head, 930 sec, 512 bytes/sect x 585937500 sectors
> sd1: sync (6.25ns offset 127), 16-bit (320.000MB/s) transfers, tagged
> queueing
> 
> Heres a response from HP tech support...
> 
> Quote:
> It is uncertain if the adaptec driver for NetBSD can handle the re-mapping
> or handle a burst of re-mappings.  It is suspect that the NetBSD adaptec
> driver is having trouble remapping unreadable hard drive blocks.  The driver

It's not a RAID controller, so I don't know why it should be the adapter's
job to remap bad sectors. The drive could do it in the case of a recovered
error, but it seems our scsictl command doesn't know how to change the
mode page for this.
If there are only a few bad blocks you can try remapping them using
scsictl sd0 reassign ... The block number is the "INFO FIELD"
in the kernel error message.

-- 
Manuel Bouyer <bouyer@antioche.eu.org>
     NetBSD: 26 ans d'experience feront toujours la difference
--