SUMMARY: Disk Problems? Help!

From: Chander Ganesan <C_at_Asu.Edu>
Date: Thu, 12 Dec 1996 21:39:29 -0700 (MST)

Thanks for the responses. I guess I'm out of luck (although no problems
with the disk since) . I'm backing up everything more regularly now and
will look into a new disk....

Thanks to the following respondants:

 "WHITTAKER, Bruce" <bjw_at_ansto.gov.au>
 alan_at_nabeth.cxo.dec.com
 Baranyai Pal <bp_at_vma.bme.hu>
 "Robert L. McMillin" <rlm_at_syseca-us.com>

-- 
Chander Ganesan
Information Technology
Arizona State University
C_at_asu.edu
Original Question:
------------
Hello...
	I've got what seems to be an odd problem to me.  Yesterday, my
messages log started filling up with the errors listed below.  I assumed
that the problem was related to my external 2 gig disk, as it was
constantly working away, even when the system was idle.  
	Anyways, I attempted to shut down the system, no dice.  The
console said that it was shutting down to single user, but X was still on
the screen, and as far as I could tell the display manager was still
running.  Eventually I got fed up with it and did the one thing I never
do...I shut off the machine.(ouch, please don't yell at me...I tried
*everything* sensible before doing so) .
	Anyways, I took it down, recycled the power and now everthing
seems to be working fine...no problems since.  Should I be expecting this
disk to take a dump on me (ie. should I pursue buying another) or is this
some strange problem that I'll only see once and is indicative of a OS
programmer with a crude sense of humor? 
Responses, as always, will be summarized....
Dec  4 18:16:54 Hornet vmunix: Defering I/O (errno 5) for block(0x6490,
0x6490) 
---------------
From: "Robert L. McMillin" <rlm_at_syseca-us.com>
Presume a failure is imminent and get a new hard drive, just in case.
I've seen these on other systems, and they can get nasty.
From: Baranyai Pal <bp_at_vma.bme.hu>
Try:
uerf -R | more
to look errlog messages. Maybe it tells you much more about disk.
> Dec  4 18:16:47 Hornet vmunix: Defering I/O (errno 5) for block(0xab920,
> 0xab920
> ) on device 8,3079
I guess that "device 8,3079" is /dev/rz3h. 8 is device major and 3079 minor
number. You can check this by:
ls -l /dev | grep "8,3079"
I found this in our /var/adm/messages:
Sep 15 19:56:31 lemma vmunix: I/O error (errno 5) for block(0x290, 0x290) on
dev 14,2
This means that read-write error occured on /dev/fd0c (floppy drive, dev. maj.
number: 14, min. number 2) caused by media error of floppy disk.
So, I think you should change the defective disk (SCSI id 3 ?).
Best wishes,
Pal Baranyai
From: alan_at_nabeth.cxo.dec.com
	errno #5 is I/O error.  The error suggests that some component
	of the kernel is going to put off writing a block (reads would
	simply have to fail) because of I/O errors.  I'd assume that
	the block numbers are LBN numbers where the write failed.  Given
	some of the wide ranges between blocks, I'd take a close look
	at the system error log (uerf or dia commands) and see where
	the errors are and what they are.  When using uerf for SCSI
	disks use the option "-o full" get the full error text.
From: "WHITTAKER, Bruce" <bjw_at_ansto.gov.au>
Hi Chander,
I would more suspect that someone had mounted a directory to a different
machine and that that machine had been rebooted or has gone down for some
reason (or even network problems). Thus the I/O is deferrred rather than
this being an I/O error. 
It may come back if the same event happens (assuming of course that is
what had happened 8-) 

Received on Fri Dec 13 1996 - 05:48:12 NZDT