Board index Solaris

Ufsdump, Write error, SCSI timeout

Ufsdump, Write error, SCSI timeout

Post by Eric L. Scho » Fri, 09 Jun 1995 04:00:00



I am having troubles with the SCSI tape device.  When I do ufsdump, I
CONSISTANTLY get write errors (see at job output below).  However, the
output of /var/adm/messages show that SCSI bus timeouts are occuring.
How can I change the SCSI bus timeout?  Will this solve the write errors?

Thanks

--------------------------  /var/adm/messages  ---------------------------
Jun  7 22:11:57 enterprise unix: WARNING: /io-unit@f,e0200000/sbi@0,0/dma@3,8100
0/esp@3,80000 (esp2):  
Jun  7 22:11:57 enterprise unix:  Connected command timeout for Target 5.0
Jun  7 22:11:57 enterprise unix: WARNING: /io-unit@f,e0200000/sbi@0,0/dma@3,8100
0/esp@3,80000/st@5,0 (st19):  
Jun  7 22:11:57 enterprise unix:  SCSI transport failed: reason 'timeout': givin
g u
Jun  7 22:11:57 enterprise unix: p  
Jun  7 22:19:47 enterprise unix: WARNING: /io-unit@f,e0200000/sbi@0,0/dma@3,8100
0/esp@3,80000 (esp2):  
Jun  7 22:19:47 enterprise unix:  Connected command timeout for Target 5.0
Jun  7 22:19:47 enterprise unix: WARNING: /io-unit@f,e0200000/sbi@0,0/dma@3,8100
0/esp@3,80000/st@5,0 (st19):  
Jun  7 22:19:47 enterprise unix:  SCSI transport failed: reason 'timeout': givin
g u
Jun  7 22:19:47 enterprise unix: p  
Jun  7 22:29:57 enterprise unix: WARNING: /io-unit@f,e0200000/sbi@0,0/dma@3,8100
0/esp@3,80000 (esp2):  
Jun  7 22:29:57 enterprise unix:  Connected command timeout for Target 5.0
Jun  7 22:29:57 enterprise unix: WARNING: /io-unit@f,e0200000/sbi@0,0/dma@3,8100
0/esp@3,80000/st@5,0 (st19):  
Jun  7 22:29:57 enterprise unix:  SCSI transport failed: reason 'timeout': givin
g u
Jun  7 22:29:57 enterprise unix: p  
Jun  7 22:35:07 enterprise unix: WARNING: /io-unit@f,e0200000/sbi@0,0/dma@3,8100
0/esp@3,80000 (esp2):  
Jun  7 22:35:07 enterprise unix:  Connected command timeout for Target 5.0
Jun  7 22:35:07 enterprise unix: WARNING: /io-unit@f,e0200000/sbi@0,0/dma@3,8100
0/esp@3,80000/st@5,0 (st19):  
Jun  7 22:35:07 enterprise unix:  SCSI transport failed: reason 'timeout': givin
g u
Jun  7 22:35:07 enterprise unix: p  
Jun  7 23:02:27 enterprise unix: WARNING: /io-unit@f,e0200000/sbi@0,0/dma@3,8100
0/esp@3,80000 (esp2):  
Jun  7 23:02:27 enterprise unix:  Connected command timeout for Target 5.0
Jun  7 23:02:28 enterprise unix: WARNING: /io-unit@f,e0200000/sbi@0,0/dma@3,8100
0/esp@3,80000/st@5,0 (st19):  
Jun  7 23:02:28 enterprise unix:  SCSI transport failed: reason 'timeout': givin
g u

----------------------------  AT job output  -----------------------------

Your "at" job "/var/spool/cron/atjobs/802576800.a" produced the following output:

Starting backup at Wed Jun 7 22:00:00 EDT 1995
mt -f $Tape rewind
ufsdump 0f $Tape /dev/rdsk/c0t0d0s0 #  /                  
  DUMP: Writing 32 Kilobyte records
  DUMP: Date of this level 0 dump: Wed Jun 07 22:00:02 1995
  DUMP: Date of last level 0 dump: the epoch
  DUMP: Dumping /dev/rdsk/c0t0d0s0 (enterprise:/) to /dev/rmt/2un.
  DUMP: Mapping (Pass I) [regular files]
  DUMP: Mapping (Pass II) [directories]
  DUMP: Estimated 30898 blocks (15.09MB).
  DUMP: Dumping (Pass III) [directories]
  DUMP: Dumping (Pass IV) [regular files]
  DUMP: 30846 blocks (15.06MB) on 1 volume at 245 KB/sec
  DUMP: DUMP IS DONE
ufsdump 0f $Tape /dev/rdsk/c0t0d0s5 #  /var              
  DUMP: Writing 32 Kilobyte records
  DUMP: Date of this level 0 dump: Wed Jun 07 22:01:09 1995
  DUMP: Date of last level 0 dump: the epoch
  DUMP: Dumping /dev/rdsk/c0t0d0s5 (enterprise:/var) to /dev/rmt/2un.
  DUMP: Mapping (Pass I) [regular files]
  DUMP: Mapping (Pass II) [directories]
  DUMP: Estimated 163826 blocks (79.99MB).
  DUMP: Dumping (Pass III) [directories]
  DUMP: Dumping (Pass IV) [regular files]
  DUMP: 163710 blocks (79.94MB) on 1 volume at 656 KB/sec
  DUMP: DUMP IS DONE
ufsdump 0f $Tape /dev/rdsk/c0t0d0s6 #  /usr              
  DUMP: Writing 32 Kilobyte records
  DUMP: Date of this level 0 dump: Wed Jun 07 22:03:17 1995
  DUMP: Date of last level 0 dump: the epoch
  DUMP: Dumping /dev/rdsk/c0t0d0s6 (enterprise:/usr) to /dev/rmt/2un.
  DUMP: Mapping (Pass I) [regular files]
  DUMP: Mapping (Pass II) [directories]
  DUMP: Estimated 375510 blocks (183.35MB).
  DUMP: Dumping (Pass III) [directories]
  DUMP: Dumping (Pass IV) [regular files]
  DUMP: 375486 blocks (183.34MB) on 1 volume at 627 KB/sec
  DUMP: DUMP IS DONE
ufsdump 0f $Tape /dev/rdsk/c0t3d0s7 #  /opt              
  DUMP: Writing 32 Kilobyte records
  DUMP: Date of this level 0 dump: Wed Jun 07 22:08:23 1995
  DUMP: Date of last level 0 dump: the epoch
  DUMP: Dumping /dev/rdsk/c0t3d0s7 (enterprise:/opt) to /dev/rmt/2un.
  DUMP: Mapping (Pass I) [regular files]
  DUMP: Mapping (Pass II) [directories]
  DUMP: Estimated 829922 blocks (405.24MB).
  DUMP: Dumping (Pass III) [directories]
  DUMP: Dumping (Pass IV) [regular files]
  DUMP: Write error 776 feet into tape 1
  DUMP: NEEDS ATTENTION: Do you want to restart?: ("yes" or "no")    DUMP: The ENTIRE dump is aborted.
ufsdump 0f $Tape /dev/rdsk/c0t2d0s7 #  /local/software    
  DUMP: Writing 32 Kilobyte records
  DUMP: Date of this level 0 dump: Wed Jun 07 22:12:01 1995
  DUMP: Date of last level 0 dump: the epoch
  DUMP: Dumping /dev/rdsk/c0t2d0s7 (enterprise:/local/software) to /dev/rmt/2un.
  DUMP: Mapping (Pass I) [regular files]
  DUMP: Mapping (Pass II) [directories]
  DUMP: Estimated 861950 blocks (420.87MB).
  DUMP: Dumping (Pass III) [directories]
  DUMP: Dumping (Pass IV) [regular files]
  DUMP: Write error 2503 feet into tape 1
  DUMP: NEEDS ATTENTION: Do you want to restart?: ("yes" or "no")    DUMP: The ENTIRE dump is aborted.
ufsdump 0f $Tape /dev/rdsk/c0t1d0s7 #  /local/enterprise_disc1
  DUMP: Writing 32 Kilobyte records
  DUMP: Date of this level 0 dump: Wed Jun 07 22:19:51 1995
  DUMP: Date of last level 0 dump: the epoch
  DUMP: Dumping /dev/rdsk/c0t1d0s7 (enterprise:/local/enterprise_disc1) to /dev/rmt/2un.
  DUMP: Mapping (Pass I) [regular files]
  DUMP: Mapping (Pass II) [directories]
  DUMP: Estimated 796852 blocks (389.09MB).
  DUMP: Dumping (Pass III) [directories]
  DUMP: Dumping (Pass IV) [regular files]
  DUMP: Write error 4733 feet into tape 1
  DUMP: NEEDS ATTENTION: Do you want to restart?: ("yes" or "no")    DUMP: The ENTIRE dump is aborted.
ufsdump 0f $Tape /dev/rdsk/c1t0d0s7 #  /local/enterprise_disc4
  DUMP: Writing 32 Kilobyte records
  DUMP: Date of this level 0 dump: Wed Jun 07 22:30:01 1995
  DUMP: Date of last level 0 dump: the epoch
  DUMP: Dumping /dev/rdsk/c1t0d0s7 (enterprise:/local/enterprise_disc4) to /dev/rmt/2un.
  DUMP: Mapping (Pass I) [regular files]
  DUMP: Mapping (Pass II) [directories]
  DUMP: Estimated 712296 blocks (347.80MB).
  DUMP: Dumping (Pass III) [directories]
  DUMP: Dumping (Pass IV) [regular files]
  DUMP: Write error 1007 feet into tape 1
  DUMP: NEEDS ATTENTION: Do you want to restart?: ("yes" or "no")    DUMP: The ENTIRE dump is aborted.
ufsdump 0f $Tape /dev/rdsk/c1t1d0s7 #  /local/enterprise_disc5
  DUMP: Writing 32 Kilobyte records
  DUMP: Date of this level 0 dump: Wed Jun 07 22:35:08 1995
  DUMP: Date of last level 0 dump: the epoch
  DUMP: Dumping /dev/rdsk/c1t1d0s7 (enterprise:/local/enterprise_disc5) to /dev/rmt/2un.
  DUMP: Mapping (Pass I) [regular files]
  DUMP: Mapping (Pass II) [directories]
  DUMP: Estimated 848872 blocks (414.49MB).
  DUMP: Dumping (Pass III) [directories]
  DUMP: Dumping (Pass IV) [regular files]
  DUMP: 83.96% done, finished in 0:01
  DUMP: 848830 blocks (414.47MB) on 1 volume at 588 KB/sec
  DUMP: DUMP IS DONE
ufsdump 0f $Tape /dev/rdsk/c1t2d0s7 #  /local/enterprise_disc6
  DUMP: Writing 32 Kilobyte records
  DUMP: Date of this level 0 dump: Wed Jun 07 22:48:18 1995
  DUMP: Date of last level 0 dump: the epoch
  DUMP: Dumping /dev/rdsk/c1t2d0s7 (enterprise:/local/enterprise_disc6) to /dev/rmt/2un.
  DUMP: Mapping (Pass I) [regular files]
  DUMP: Mapping (Pass II) [directories]
  DUMP: Estimated 735656 blocks (359.21MB).
  DUMP: Dumping (Pass III) [directories]
  DUMP: Dumping (Pass IV) [regular files]
  DUMP: 735614 blocks (359.19MB) on 1 volume at 726 KB/sec
  DUMP: DUMP IS DONE
ufsdump 0f $Tape /dev/rdsk/c1t3d0s7 #  /local/enterprise_disc7
  DUMP: Writing 32 Kilobyte records
  DUMP: Date of this level 0 dump: Wed Jun 07 22:56:52 1995
  DUMP: Date of last level 0 dump: the epoch
  DUMP: Dumping /dev/rdsk/c1t3d0s7 (enterprise:/local/enterprise_disc7) to /dev/rmt/2un.
  DUMP: Mapping (Pass I) [regular files]
  DUMP: Mapping (Pass II) [directories]
  DUMP: Estimated 899464 blocks (439.19MB).
  DUMP: Dumping (Pass III) [directories]
  DUMP: Dumping (Pass IV) [regular files]
  DUMP: Write error 2333 feet into tape 1
  DUMP: NEEDS ATTENTION: Do you want to restart?: ("yes" or "no")    DUMP: The ENTIRE dump is aborted.
mt -f $Tape rewoffl
/dev/rmt/2un: Device busy
date
Wed Jun  7 23:02:28 EDT 1995

 
 
 
Top

Ufsdump, Write error, SCSI timeout

Post by Patrick D. McSwigg » Fri, 09 Jun 1995 04:00:00




> I am having troubles with the SCSI tape device.  When I do ufsdump, I
> CONSISTANTLY get write errors (see at job output below).  However, the
> output of /var/adm/messages show that SCSI bus timeouts are occuring.
> How can I change the SCSI bus timeout?  Will this solve the write errors?

> Thanks

I was seeing a similar problem with our Sun tape drive. I don't recall
what /var/adm/messages showed, but the dump output was the same. The
lights on the drive would indicate a "high soft error" (whatever that is).
Once an error occurred both the drive and tape were unusable. Running the
cleaning cassette would seem to "reset" the drive, but if the tape was to
be reused it had to be explicitely erased. This could happen as much as
every other dump.

The drive was brand new and Sun kept shipping us new parts. (Not counting
the case, there are only three.) By the time they were done, the drive
mech. had been replaced twice and everything else once. They never did
have an explaination for what was wrong.

The problem has not gone away completely, but keeping the heads super
clean seems to help. I run an 8 tape dump cycle, and I now run the
cleaning tape just before the start of each new cycle. That amounts to
probably 5 hrs total use between cleanings -- *much* less than the
recommended 20-25hrs.

I wish I knew what the problem was, but Sun tech support had absolutely
*no* suggestions on how I could track down the problem, including how to
find out what sort of "high soft error" it was.

--


University of Cincinnati

 
 
 
Top

Ufsdump, Write error, SCSI timeout

Post by J.A. Pete » Fri, 09 Jun 1995 04:00:00



> I am having troubles with the SCSI tape device.  When I do ufsdump, I
> CONSISTANTLY get write errors (see at job output below).  However, the
> output of /var/adm/messages show that SCSI bus timeouts are occuring.
> How can I change the SCSI bus timeout?  Will this solve the write errors?
> --------------------------  /var/adm/messages  ---------------------------


> Jun  7 22:11:57 enterprise unix:  Connected command timeout for Target 5.0


> Jun  7 22:11:57 enterprise unix:  SCSI transport failed: reason 'timeout':

...<snip>...

You either have a hardware or cabling fault. The hardware failure could
be caused by the SCSI controller or most likely the tape drive. Debugging
the cabling and termination is not much of trick. I would recommend a
call to your hardware maintenance provider.

Cheer,s

t

 
 
 
Top

Ufsdump, Write error, SCSI timeout

Post by Dr Ivan S. Bisho » Sat, 10 Jun 1995 04:00:00


I'd make sure that you keep the block size 'sensible'. I have seen
problems when writing records over 64K (128 blocks). This may be
because I have misssed a patch of course, BUT try fixing the
block size to <= 128 and see what happens. Hope it works for you.

Ivan S. Bishop

 
 
 
Top

Ufsdump, Write error, SCSI timeout

Post by Andre Be » Sun, 11 Jun 1995 04:00:00



|> I am having troubles with the SCSI tape device.  When I do ufsdump, I
|> CONSISTANTLY get write errors (see at job output below).  However, the
|> output of /var/adm/messages show that SCSI bus timeouts are occuring.
|> How can I change the SCSI bus timeout?  Will this solve the write errors?

How did you configure the SCSI Tape ? Is it one of the 'supported by Sun'
Models ? If not, you may need to configure it in /kernel/drv/st.conf.
Do a man st for more information on how to do this. Especially, use the
flag mentioned there for long timeouts. And, if on 2.3, _ignore_ the phrase
telling you that giving a block size of 0 would be valid for variable BS
tapes - it will not work. Use 512 or 1024 (is there an upper limit ? I once
used 32768 and could write perfectly, but the machine instantly paniced on
any try of reading...).

--
+-o-+--------------------------------------------------------+-o-+
| o |               \\\- Brain Inside -///                   | o |
| o |                   ^^^^^^^^^^^^^^                       | o |

+-o-+--------------------------------------------------------+-o-+

 
 
 
Top

1. Ufsdump, Write error, SCSI timeout: SOLVED

I previously reported a problem where I was obtaining write errors while
running ufsdump.  The /var/adm/message file showed SCSI timeouts were
occurring.

Experimentation showed that
  1.  The errors occurred on 8mm also with tar and also while reading.
  2.  I did not see the timeout errors with a 4mm drive.

I installed patch 102002-03.  This solved the problem.  No tape drive
timeouts.

+-------------------------------------------------------------------------+
|    Eric L. Schott          "And when you get lost in life,  know that   |

|    (814) 863-4194           Albert Schweitzer,    letter to god-child   |
+-------------------------------------------------------------------------+

2. Meta and Alt keys on KDE

3. SCSI transport error: timeout waiting to write packet

4. cron dying - HELP!

5. solaris 2.8, disksuite 4.2.1, ufsdump -- write errors

6. swap partition has shrunk!

7. ufsdump write error

8. Linux trashes windows

9. UFSDUMP: Write error 501 feet into tape 1

10. HELP! WARNING & write errors on ufsdump.

11. scsi timeouts while writing cdrom

12. Problems with NCR/RZ25L setup and SCSI Write timeouts on RH5.1 Linux


All times are UTC
Board index