A few weeks ago, I asked for help concerning intermittent read/write errors on
a Fujitsu M2263SA attached to 4/330:
>One of the Fujitsu M2263SA SCSI disks on our Sun 4/330 running SunOS 4.1.1
>has recently been acting up. Twice in the last month this drive as apparently
>gone offline after an I/O failure. A few questions:
>
>1) Does anyone know what hardware error code 0x44 and 0xc4 are, for this
> drive? The installation/specification OEM manual I have for this drive
> doesn't have a list of error codes (it says to see the CE manual, which I
> don't have). These are the errors reported by the drive (see below).
>
>2) According to the 4.1.1 release manual, p. 24, the 4/330 is capable of
> synchronous SCSI transmission. However, unlike our SS2's, the 4/330
> doesn't report that "target n now synchronous a xx.x mb/s ..." at boot
> time. Is there something odd about the SCSI controller on the 4/330 and
> 4.1.1 which may be causing the problem? (The controller in question is
> sm0 on the CPU board). The SCSI bus is well under 6 meters. The only
> odd thing about the SCSI bus is that the last device on the chain, which
> has an external terminator on it, is kept powered down (it's an M/O drive
> whose driver doesn't work with 4.1.1--we'll be moving it to another
> machine in the near future--for now, it's only purpose is to terminate
> the SCSI bus). BTW, what is the SCSI level of the sm controller on the
> 4/330, SCSI-1 or 2?
>
> ...
>
> Mar 5 14:13:32 vmunix: sd2c: write retry, block 150592 (150592 relative)
> Mar 5 14:13:32 vmunix: sd2c error: sense key(0x4): hardware error, error
> code(0x44):
> ...
Unfortunately, I didn't receive any definitive answers to my questions (very
rare for this list!). Most suggested using format to fix the bad blocks
reported by the kernel. However, since the analyze subcommand of format
didn't find any errors in and around the blocks in question, plus the fact
that the drive would remain unusable after the read/write failure until
power-cycled (I've encountered bad blocks before on an SMD disk--the kernel
would report the error in the same manner, but the disk would still be
usable), led be to suspect a failing controller on the disk. Another write
error occurred last week, and again, the drive was unusable until power-
cycled. I have since replaced the drive. In perhaps a related problem, the
power supply in the enclosure containing this drive (also containing another
fujitsu 2263 and an exabyte) completely failed last night. Perhaps the
failures on the drive were due to power being dropped by the power supply on
the line to the drive; hence the drive would go offline because power was no
longer being supplied to it. Actually, the more I think about this last
explanation, the more plausible it seems.
In regards to my second set of questions, the 4/330 does indeed use sync-SCSI
under 4.1.1 and greater (as does, I assume, any 4/300 based sun4). This was
confirmed by Sun.
Thanks to all who replied:
Mike Raffety <miker@sbcoc.com>
Eckhard R"uggeberg <eckhard@ts.go.dlr.de>
Cuong C Nguyen <cuongc@nad.3com.com>
Patrick R. Demorais <demorais@ssd.Kodak.Com>
John Valdes Department of the Geophysical Sciences
valdes@geosun.uchicago.edu University of Chicago
This archive was generated by hypermail 2.1.2 : Fri Sep 28 2001 - 23:06:41 CDT