ahd0: Invalid Sequencer interrupt occurred.
Ade Lovett
ade at FreeBSD.org
Thu Nov 10 18:05:40 PST 2005
On Nov 10, 2005, at 05:30 , Ricardo A. Reis wrote:
Reducing the problem to the relevant pieces:
> ahd0: <Adaptec AIC7902 Ultra320 SCSI adapter> port 0x2400-0x24ff,
> 0x2000-0x20ff mem 0xdd200000-0xdd201fff irq 32 at device 2.0 on pci3
> ahd0: [GIANT-LOCKED]
> aic7902: Ultra320 Wide Channel A, SCSI Id=7, PCI-X 67-100Mhz, 512 SCBs
> ahd1: <Adaptec AIC7902 Ultra320 SCSI adapter> port 0x2c00-0x2cff,
> 0x2800-0x28ff mem 0xdd202000-0xdd203fff irq 33 at device 2.1 on pci3
> ahd1: [GIANT-LOCKED]
> aic7902: Ultra320 Wide Channel B, SCSI Id=7, PCI-X 67-100Mhz, 512 SCBs
[...]
> da0 at ahd0 bus 0 target 0 lun 0
> da0: <SEAGATE ST373307LC 0006> Fixed Direct Access SCSI-3 device
> da0: 320.000MB/s transfers (160.000MHz, offset 63, 16bit), Tagged
> Queueing Enabled
> da0: 70007MB (143374744 512 byte sectors: 255H 63S/T 8924C)
> da1 at ahd0 bus 0 target 1 lun 0
> da1: <SEAGATE ST373307LC 0006> Fixed Direct Access SCSI-3 device
> da1: 320.000MB/s transfers (160.000MHz, offset 63, 16bit), Tagged
> Queueing Enabled
> da1: 70007MB (143374744 512 byte sectors: 255H 63S/T 8924C)
> Trying to mount root from ufs:/dev/da0s1a
Adaptec HBAs and Seagate drives have a long and intensely painful
history of not working well together. Adaptec blames Seagate.
Seagate blames Adaptec. Throw in the myriad of subtly different AIC
controllers that are commonplace on 1U and 2U rackmount servers, and
things get even more entertaining.
You essentially have 3 options
1) replace the HBA -- somewhat difficult to do if it's embedded and
you need the PCIX slots for something else.
2) replace the drives -- IBM/Hitachi are fine choices here. Make
sure to tell whomever you purchase systems from that you'll not
accept Seagate drives in the future.
3) inside the adaptec bios, drop the drives to U160 speed, making
sure that *both* packetizing *and* QAS are turned OFF. You'll lose a
little bit of performance (but not all that much, Seagate drives
really are garbage), and get some semblance of stability.
-aDe
More information about the freebsd-stable
mailing list