Discussion:
Corrupted CCKD
(too old to reply)
holloway_tim
2005-09-23 11:54:59 UTC
Permalink
Someone played games with the UPS's and it appears to have damaged
some of my DASD volumes - and/or this comes from the known geometry
problem in the version of Tur(n)key that the volume was initialzed from.

Anyway:

[***@here pds]# cckd2ckd ../dasd/pub000.240 pub000.240
HHCDA004I opening ../dasd/pub000.240 readonly
HHCDA020I ../dasd/pub000.240 cyls=555 heads=30 tracks=16650 trklen=19456
cckdcdsk: Recoverable header errors found: 11427243 9405715 8315964
1089751 1089
440 311 1
cckdcdsk: forcing check level 1
cckdcdsk: free space errors found: file not closed
cckdcdsk: errors detected on read-only file
HHCDU044I Creating 3350 volume : 555 cyls, 30 trks/cyl, 19456 bytes/track
HHCDU041I 555 cylinders successfully written to file pub000.240
HHCDA020I pub000.240 cyls=0 heads=30 tracks=0 trklen=19456
0% 0 of 16650HHCDA032E error reading trk 0: lseek error:
Illegal seek
HHCDC009E cckd2ckd: pub000.240 write error track 0 stat=0E
[***@finlux pds]# ls -lh
total 12K
-rw-r--r-- 1 root root 0 Sep 22 16:40 algol.mac
-rw-r----- 1 root root 512 Sep 22 16:48 pub000.240
[***@here pds]#

I get the impression that when Hercules starts up, it runs cckdcdsk's
on the volumes and it's supposed (?) to repair any damage that might
have come from a crash. However, it doesn't look repaired, and I'm not
sure what that "read-only" is supposed to mean here, but Hercules IS
running while I did this.

And for that matter, what's the WRITE error? This is a Linux LVM
volume and it's only 35% full.

At this stage, I'd be happy if I could just recover as many datasets
as I could off this volume and build a new volume for them, but any
and all suggestions are welcome!

Tim




------------------------ Yahoo! Groups Sponsor --------------------~-->
Fair play? Video games influencing politics. Click and talk back!
http://us.click.yahoo.com/T8sf5C/tzNLAA/TtwFAA/W4wwlB/TM
--------------------------------------------------------------------~->
Enchanter.tim
2005-09-23 12:01:43 UTC
Permalink
Tim,
Judging from the ls output, the "disk" is 512 bytes in size. It looks to me
that the disk is..umm..gone.

512 bytes is a bit small

Kind Regards,

Enchanter_tim
(The Other Tim)

-----Original Message-----
From: hercules-390-***@public.gmane.org [mailto:hercules-390-***@public.gmane.org] On
Behalf Of holloway_tim
Sent: vrijdag 23 september 2005 13:55
To: hercules-390-***@public.gmane.org
Subject: [hercules-390] Corrupted CCKD

Someone played games with the UPS's and it appears to have damaged some of
my DASD volumes - and/or this comes from the known geometry problem in the
version of Tur(n)key that the volume was initialzed from.

Anyway:

[***@here pds]# cckd2ckd ../dasd/pub000.240 pub000.240 HHCDA004I opening
../dasd/pub000.240 readonly HHCDA020I ../dasd/pub000.240 cyls=555 heads=30
tracks=16650 trklen=19456
cckdcdsk: Recoverable header errors found: 11427243 9405715 8315964
1089751 1089
440 311 1
cckdcdsk: forcing check level 1
cckdcdsk: free space errors found: file not closed
cckdcdsk: errors detected on read-only file HHCDU044I Creating 3350 volume :
555 cyls, 30 trks/cyl, 19456 bytes/track HHCDU041I 555 cylinders
successfully written to file pub000.240 HHCDA020I pub000.240 cyls=0 heads=30
tracks=0 trklen=19456
0% 0 of 16650HHCDA032E error reading trk 0: lseek error:
Illegal seek
HHCDC009E cckd2ckd: pub000.240 write error track 0 stat=0E [***@finlux
pds]# ls -lh total 12K
-rw-r--r-- 1 root root 0 Sep 22 16:40 algol.mac
-rw-r----- 1 root root 512 Sep 22 16:48 pub000.240 [***@here pds]#

I get the impression that when Hercules starts up, it runs cckdcdsk's on the
volumes and it's supposed (?) to repair any damage that might have come from
a crash. However, it doesn't look repaired, and I'm not sure what that
"read-only" is supposed to mean here, but Hercules IS running while I did
this.

And for that matter, what's the WRITE error? This is a Linux LVM volume and
it's only 35% full.

At this stage, I'd be happy if I could just recover as many datasets as I
could off this volume and build a new volume for them, but any and all
suggestions are welcome!

Tim







------------------------ Yahoo! Groups Sponsor --------------------~-->
Fair play? Video games influencing politics. Click and talk back!
http://us.click.yahoo.com/T8sf5C/tzNLAA/TtwFAA/W4wwlB/TM
--------------------------------------------------------------------~->
g***@public.gmane.org
2005-09-23 12:53:03 UTC
Permalink
Try running cckdcdsk first

cckdcdsk -3 ../dasd/pub000.240

Greg
Post by holloway_tim
HHCDA004I opening ../dasd/pub000.240 readonly
HHCDA020I ../dasd/pub000.240 cyls=555 heads=30 tracks=16650
trklen=19456cckdcdsk: Recoverable header errors found: 11427243
9405715 8315964
1089751 1089
440 311 1
cckdcdsk: forcing check level 1
cckdcdsk: free space errors found: file not closed
cckdcdsk: errors detected on read-only file
------------------------ Yahoo! Groups Sponsor --------------------~-->
Fair play? Video games influencing politics. Click and talk back!
http://us.click.yahoo.com/T8sf5C/tzNLAA/TtwFAA/W4wwlB/TM
--------------------------------------------------------------------~->
Peter Vels
2005-09-24 00:40:43 UTC
Permalink
Post by holloway_tim
HHCDA004I opening ../dasd/pub000.240 readonly
^^^^^^^^
Post by holloway_tim
cckdcdsk: errors detected on read-only file
^^^^^^^^^
Why is pub000.240 read-only? Surely if cckdcdsk only has read-only access
to pu000.240 then it won't be able to repair it.

I had a similar problem a while back. I was using shadow files (which were
fine) on top of read-only volumes which had errors. The "solution" was just
to merge the shadow files. The resultant merged files were then error-free.
I made them readonly again and created new shadow files.

Cheers,
Peter


------------------------ Yahoo! Groups Sponsor --------------------~-->
Get Bzzzy! (real tools to help you find a job). Welcome to the Sweet Life.
http://us.click.yahoo.com/A77XvD/vlQLAA/TtwFAA/W4wwlB/TM
--------------------------------------------------------------------~->
Greg Smith
2005-09-24 01:21:48 UTC
Permalink
Post by Peter Vels
Post by holloway_tim
HHCDA004I opening ../dasd/pub000.240 readonly
^^^^^^^^
Post by holloway_tim
cckdcdsk: errors detected on read-only file
^^^^^^^^^
Why is pub000.240 read-only? Surely if cckdcdsk only has read-only access
to pu000.240 then it won't be able to repair it.
I had a similar problem a while back. I was using shadow files (which were
fine) on top of read-only volumes which had errors. The "solution" was just
to merge the shadow files. The resultant merged files were then error-free.
I made them readonly again and created new shadow files.
Umm, on cckd2ckd the in file is opened read-only. In other words,
dasdcopy (and variants including cckd2ckd) assume an undamaged in file.

For the similar problem, I do recall a situation that Volker came
across. It seems that if the base file is specified to be read-only it
is opened read-only but if hercules was running under root you could do
a backwards merge. In this case, you now have to do (I think) `sf-xxxx
force'.

I'm open to change my views given a convincing argument. One argument
for the status quo is that dasdcopy doesn't change the in file.

Greg



------------------------ Yahoo! Groups Sponsor --------------------~-->
Get Bzzzy! (real tools to help you find a job). Welcome to the Sweet Life.
http://us.click.yahoo.com/A77XvD/vlQLAA/TtwFAA/W4wwlB/TM
--------------------------------------------------------------------~->
holloway_tim
2005-09-24 11:16:27 UTC
Permalink
Post by Greg Smith
Post by Peter Vels
Post by holloway_tim
HHCDA004I opening ../dasd/pub000.240 readonly
^^^^^^^^
Post by holloway_tim
cckdcdsk: errors detected on read-only file
^^^^^^^^^
Why is pub000.240 read-only? Surely if cckdcdsk only has
read-only access
Post by Greg Smith
Post by Peter Vels
to pu000.240 then it won't be able to repair it.
I had a similar problem a while back. I was using shadow files (which were
fine) on top of read-only volumes which had errors. The
"solution" was just
Post by Greg Smith
Post by Peter Vels
to merge the shadow files. The resultant merged files were then error-free.
I made them readonly again and created new shadow files.
Umm, on cckd2ckd the in file is opened read-only. In other words,
dasdcopy (and variants including cckd2ckd) assume an undamaged in file.
For the similar problem, I do recall a situation that Volker came
across. It seems that if the base file is specified to be read-only it
is opened read-only but if hercules was running under root you could do
a backwards merge. In this case, you now have to do (I think) `sf-xxxx
force'.
I'm open to change my views given a convincing argument. One argument
for the status quo is that dasdcopy doesn't change the in file.
Greg
Right after I posted, I discovered the "cckdcksk -3 -f" option.
Although it seems to have cleaned up SOME of the problem, my full-pack
copy exercise still fails:
[***@oraserver dasd]$ cckd2ckd pub000.240 pub000.240.ckd
HHCDA004I opening pub000.240 readonly
HHCDA020I pub000.240 cyls=555 heads=30 tracks=16650 trklen=19456
HHCDU044I Creating 3350 volume : 555 cyls, 30 trks/cyl, 19456 bytes/track
HHCDU041I 555 cylinders successfully written to file pub000.240.ckd
HHCDA020I pub000.240.ckd cyls=0 heads=30 tracks=0 trklen=19456
0% 0 of 16650HHCDA032E error reading trk 0: lseek error:
Illegal seek
HHCDC009E cckd2ckd: pub000.240.ckd write error track 0 stat=0E

Fortunately, it looks like there's nothing on this pack that I can't
reconstruct, and many of the files in it are readable (the 512 byte
output size is simply an indication that the copy operation failed).

I'm beginning to see the virtue of shadow volumes, though. On a RedHat
architecture it's possible to put together a Hercules as a system
service that gets shut down by the ups daemon, but, alas, on a P500
the time it takes to shut down MVS is too long to ensure that someone
won't yank the underlying system out in the middle of the process.

Speaking of which: I built a hercules from CVS and noticed that the
command line doesn't report "-d" as an option.

Tim





------------------------ Yahoo! Groups Sponsor --------------------~-->
Most low income households are not online. Help bridge the digital divide today!
http://us.click.yahoo.com/cd_AJB/QnQLAA/TtwFAA/W4wwlB/TM
--------------------------------------------------------------------~->
mario_bezzi
2005-09-24 13:12:35 UTC
Permalink
Speaking of shadow files, which I find great, I would like to have an
option to always start an Hercules work session with a new set
of 'temporary' shadow files (one for each DASD volume) and have them
merged or discarded at the end of the session. VMWare had (has?)
something similar and I think it would not be overly difficult to
implement it in Hercules. It would also protect emulated DASD files
from sudden host system's crashes.

just a thought..

mario




------------------------ Yahoo! Groups Sponsor --------------------~-->
Get Bzzzy! (real tools to help you find a job). Welcome to the Sweet Life.
http://us.click.yahoo.com/A77XvD/vlQLAA/TtwFAA/W4wwlB/TM
--------------------------------------------------------------------~->
John Preston
2005-09-24 13:28:56 UTC
Permalink
You can do that easily via whatever batch file or script you use to start Hercules. No need for it to be added to the product.

One method is to make a copy during startup to a new directory of all the shadow files at that point in time, as a fallback. Another way is to run multiple nested levels of shadow files. VMWare is nowhere near as easy to use as the Hercules shadow file system in this way.

I feel that shadow file merging should not be automatic, rather only done on demand (and I do it fairly rarely). I do that with a different command (batch/script) file.

The primary DASD files therefore are very protected, and of course you have backups ... I shadows were automatically merged, then that would remove a degree of that protection.

Regards,
John Preston
Post by mario_bezzi
Speaking of shadow files, which I find great, I would like to have an
option to always start an Hercules work session with a new set
of 'temporary' shadow files (one for each DASD volume) and have them
merged or discarded at the end of the session. VMWare had (has?)
something similar and I think it would not be overly difficult to
implement it in Hercules. It would also protect emulated DASD files
from sudden host system's crashes.
just a thought..
mario
------------------------ Yahoo! Groups Sponsor --------------------~-->
Most low income households are not online. Help bridge the digital divide today!
http://us.click.yahoo.com/cd_AJB/QnQLAA/TtwFAA/W4wwlB/TM
--------------------------------------------------------------------~->
Fish
2005-11-20 04:33:28 UTC
Permalink
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

(trying to get caught up on my huge two-months backlog of
unread/unanswered email...)

mario_bezzi wrote:
(back on Saturday, September 24, 2005 6:13 AM)
Post by mario_bezzi
Speaking of shadow files, which I find great,
Aren't they nice though?? I really think they're great too. (They
certainly came in REAL handy during this latest Herc bug-hunt I can
tell you!) I think Greg did a fantastic job with them and we should
all thank him for his efforts on it. It was his invention. (Thanks
Greg!)

Once thing I wish was possible though (which isn't today) was the
ability to create snapshots for NON-compressed dasd. Right now you
can only create them for compressed dasd files I believe, which seems
to be an artificial limitation. Greg? Is there something I'm missing
here? Is there a reason snapshots are only supported for compressed
dasd? How tough would it be to add support for them for regular
non-compressed dasd as well?
Post by mario_bezzi
I would like to have an option to always start an Hercules work
session with a new set of 'temporary' shadow files (one for each
DASD volume) and have them merged or discarded at the end of the
session.
That is ANOTHER thing I was thinking was needed too! You can I think
alike mario. :)
Post by mario_bezzi
VMWare had (has?) something similar
Yep, and M$'s Virtual PC has it too. It calls them "Undo disks".
They're created whenever the virtual machine is started and whenever
it's shutdown it asks whether you want to save your changes or not
(i.e. whether you want to merge your snapshot (undo disk) back into
the base file or not). I think having such an ability would prove to
be *extremely* handy/valuable IMO.

I was toying with the idea of maybe trying to add some logic to
HercGUI to maybe try and provide such a feature (by simply automating
the entering of the proper commands) but it'd be nice if it was
built-in.
Post by mario_bezzi
and I think it would not be overly difficult to
implement it in Hercules. It would also protect emulated DASD files
from sudden host system's crashes.
just a thought..
And a good thought too IMO!

Maybe one of these days either Greg or I (or one of our users) will
get around to implementing it. :)

Then again, maybe some day monkeys will fly out my butt too.[1] ;-)

- --
"Fish" (David B. Trout) - fish(at)infidels.org
Fight Spam! Join CAUCE! <http://www.cauce.org/>
(Any HTML email received will be deleted unread)
PGP key fingerprints:
DH/DSS: 9F9B BAB0 BA7F C458 1A89 FE26 48F5 D7F4 C4EE 3E2A
RSA: 6B37 7110 7201 9917 9B0D 99E3 55DB 5D58 FADE 4A52

[1] Which I guess is my way of saying: "Maybe SOME day I'll actually
have the TIME to at least *try* an implement such a thing." :)


(time is in very short supply for me these days I'm afraid...)


-----BEGIN PGP SIGNATURE-----
Version: PGP 7.0.4

iQA/AwUBQ3/8l0j11/TE7j4qEQIPuACg1aG+VXwt5SMTH8ctCUw2CvxgKcQAnj42
MusQHYHyIC/foR8ifu0J7hnD
=Ty1u
-----END PGP SIGNATURE-----




------------------------ Yahoo! Groups Sponsor --------------------~-->
Fair play? Video games influencing politics. Click and talk back!
http://us.click.yahoo.com/u8TY5A/tzNLAA/yQLSAA/W4wwlB/TM
--------------------------------------------------------------------~->
Continue reading on narkive:
Loading...