[comp.sys.apollo] DSP90 w/CDC 500MB Drives: Disk Errors

jec@iuvax.cs.indiana.edu (James E. Conley) (03/23/88)

	I've recently experienced problems with disk errors on our two
servers.  This is our configuration:

	DSP-90 w/3MB
	2 MSD500 drives
	SR9.7 (did an SR9.7 invol)

	It appears for some reason that we are getting excessive disk
errors on BOTH systems on BOTH drives.  We are mainly getting DISK BLOCK
HEADER errors.  I tried the following to remedy the situation:

	(1) INVOL the bad disk getting a list of bad blocks to put
            in the bad block list. 

	(2) Adding the bad blocks.  This put us a 210 bad blocks on the
	    one disk I've tried this on.

	(3) RE-INVOLing the disk (got no errors, but it appears to reserved
            a 1 block partition or something like that-- I think the bad
            block table must have exceeded its normal boundaries or something
            like that)

	(4) Booting diskless and installing SR9.7

	(5) Trying to boot on that disk and getting SYSBOOT not found
            errors.

	I normally would suspect just a bad disk, or bad controller or even
a bad DSP-90, but for both systems to have so similar problems I am leaning
toward suspecting the operating system itself.  Has anyone else had this sort
of thing and what did you do?  I have Apollo Field Service on it now, but he
is a new person and I've always had bad luck with new people.

jen@mips.COM (Fred Jen) (03/31/88)

In article <7087@iuvax.cs.indiana.edu> jec@iuvax.UUCP (James E. Conley) writes:
>
>	I've recently experienced problems with disk errors on our two
>servers.  This is our configuration:
>
>	DSP-90 w/3MB
>	2 MSD500 drives
>	SR9.7 (did an SR9.7 invol)

You should not be using SR9.7 invol.  Somehow when Apollo release SR9.5 and newer
revisions, they unfix a bug fix.  The newer invol can not handle extended bad block 
list (bad block list longer than 200 or something close to it).  

If you use SR9.2.3's invol program to invol your disk, the problem  with your disks
should be fixed.

You should replace all your invol program with SR9.2.3 invol program.

>
>	It appears for some reason that we are getting excessive disk
>errors on BOTH systems on BOTH drives.  We are mainly getting DISK BLOCK
>HEADER errors.  I tried the following to remedy the situation:
>
>	(1) INVOL the bad disk getting a list of bad blocks to put
>            in the bad block list. 
>
>	(2) Adding the bad blocks.  This put us a 210 bad blocks on the
>	    one disk I've tried this on.

You bad block list is probably in the extended bad block list.

>
>	(3) RE-INVOLing the disk (got no errors, but it appears to reserved
>            a 1 block partition or something like that-- I think the bad
>            block table must have exceeded its normal boundaries or something
>            like that)

Invol forgets the backblocks you just put in.

>
>	(4) Booting diskless and installing SR9.7
>
>	(5) Trying to boot on that disk and getting SYSBOOT not found
>            errors.
>
>	I normally would suspect just a bad disk, or bad controller or even
>a bad DSP-90, but for both systems to have so similar problems I am leaning
>toward suspecting the operating system itself.  Has anyone else had this sort
>of thing and what did you do?  I have Apollo Field Service on it now, but he
>is a new person and I've always had bad luck with new people.

Try invol with SR9.2.3 invol and see if it works.  Good luck.

fred

-- 

-Fred Jen
UUCP: {decvax,ucbvax,ihnp4,hplabs}!decwrl!mips!jen
USPS: MIPS Computer Systems, 930 Arques, Sunnyvale, CA 94086, (408) 991-0306

jen@mips.COM (Fred Jen) (03/31/88)

I forgot to add.  Invol problem only happens on CDC500 drives.

fred
-- 

-Fred Jen
UUCP: {decvax,ucbvax,ihnp4,hplabs}!decwrl!mips!jen
USPS: MIPS Computer Systems, 930 Arques, Sunnyvale, CA 94086, (408) 991-0306

collins@nvpna1.UUCP (Donie Collins 44091) (04/01/88)

In article <1956@mips.mips.COM>, jen@mips.COM (Fred Jen) writes:
>In article <7087@iuvax.cs.indiana.edu> jec@iuvax.UUCP (James E. Conley) writes:
> >	I've recently experienced problems with disk errors on our two
> >servers.  This is our configuration:
> 
>You should not be using SR9.7 invol.Somehow when Apollo release SR9.5 and newer
>revisions,they unfix a bug fix.The newer invol can not handle extended bad  
>block list (bad block list longer than 200 or something close to it).  
>If you use SR9.2.3's invol program to invol your disk, the problem with your 
>disks should be fixed.
>You should replace all your invol program with SR9.2.3 invol program.
> 

On the January Patch tape there is a new invol for 9.7 (invol 9.7.0.2 dated
11/30/87), this one handles larger bad block lists and fixes a number
of other problems besides. It's offline invol, sau[1-8]. Patch 83, I think.

On the Feb patch tape patch 100, has a new invol (9.7.0.3) for /com/invol
and sau[6-8] invol.

Perhaps you should use these instead of the 9.2.3 invol. 

Good Luck.

-- 
Donal O'Coileain.   ..!mcvax!prle!nvpna1!collins or
                    ..!mcvax!prle!nvpna1!apolloway!collins
"I believe in the bells of the Christchurch - Ringing out for this land
 I believe in the powers that be - But they won't overpower me" A Celebration