SUMMARY: block read errors from ufsdump - SS10/Sol2.4

From: Joseph Youn (
Date: Sat Oct 14 1995 - 13:15:32 CDT

I only got one response, but it was right on target. I was finally
able to reboot the machine & had no problems backing up the drive...

So, in summary, if you get a lot of block read errors from an older SCSI
drive on a Solaris machine, turn off "tagged command queueing" and
see if that fixes the problem.



Solution ---

> From: misawa@physics.Berkeley.EDU (Shigeki Misawa)
> The most likely cause is "tagged command queueing".
> Try disabling tagged command queueing as directed in the Solaris FAQ.
> 3.28) How do I make Solaris2 use my Toshiba MK538FB and other older
> SCSI disks?
> Append this line to /etc/system and reboot:
> set scsi_options & ~0x80
> This turns off Command Queuing, which upsets the Toshiba and some
> other drives.
> In Solaris 2.4 and later you can set those options per SCSI bus. See
> isp(7) and esp(7).
> I had the same problem with a Micropolis 2217 on Solaris 2.4.
> Shigeki Misawa
> Graduate Student
> Solaris SysAdmin
> UCB Physics Department

Original message ---

> I have a pair of Micropolis 1936-21 drives on a 4-processor Sparc 10
> running Solaris 2.4, with all of the recommended patches from a couple of
> months ago (I know I should update them)... They are rated as 3 GB
> drives. They might be the A/V version of that drive if that makes any
> difference at all.
> I am using the following entry in my format.dat for the drives:
> ... lines deleted ...
> During normal use, I have no problems with the drive. However, when I try
> to back up the drive, I get errors of the following format:
> DUMP: Warning - cannot read sector xxxxxxx of `/dev/rdsk/c1t4d0s7'
> DUMP: Warning - block xxxxxxx is beyond the end of `/dev/rdsk/c1t4d0s7'
> DUMP: bread: llseek2 error
> DUMP: More than 32 block read errors from dump device
> and the dump fails. I forgot to mention that there is a Buffered
> Ehternet/SCSI card and a FDDI card in the machine. Since all of the
> SBUS slots are full or blocked by the MBUS modules, I'm running a
> VT320 for the system console...
> The sectors from the first warning appear to come from all over the
> disk.
> At first, I assumed this was a simple format error, and I brought the
> system down to single user mode. Since I couldn't run a dump, I used
> GNU tar to back the files up to another disk. I then reformatted
> the drive (It had been previously formatted with different values).
> I dumped the empty drive - that worked. I dumped the tar files from the
> othe filesystem - that worked. I untarred the files back onto the
> original drive & tried to dump them & got the same errors.
> I am absolutely lost - does anyone know of an alternative to buying new
> drives and replacing these???

This archive was generated by hypermail 2.1.2 : Fri Sep 28 2001 - 23:10:34 CDT