topic Re: question on mount verification loop in Operating System - OpenVMS

question on mount verification loop

antonio elder prado — Wed, 28 Jul 2010 01:43:32 GMT

Hi,
when i boot my ovms 8.2 (on ia64, rx2600) it goes on mountver ovms$dka100 is offline
mount verification in progress
and keeps on repeating forever never timing out. (i left it running for more than 3 hours)

i did try to ^p however it gives me on
(crash y/n), if i select y writes the dump and exits it out to rebooting with the same result,
if i select n goes back to the endless loop of offline, mount verification.

i do have a ovms 8.3 CDs, and i was able to get the $$$ prompt where i could see the IPC (by ^p) but cannnot cancel the mounting because the mounting verification is not being performed, and i am not familiar with the ovms and do not know much what to do that can break the mount verification when booting out of the disk.

can you please guide me on what to do?

thank you in advance,
./antonio/.

Re: question on mount verification loop

Hoff — Wed, 28 Jul 2010 02:04:46 GMT

That VMS is too old to have the ^P stuff; that got added with V8.2-1 release.

That mount verification behavior is usually indicative of a storage connection problem, or of a disk or SCSI controller or SCSI bus hardware error.

With this also running V8.2, this could well be software, too; there have been a number of fixes to SCSI, both with ECOs and UPDATE kits, and with the V8.3 or V8.3-1H1 releases.

Make sure the disk is seated and latched.

Swap the disk. See if the error moves with the slot.

What's the history of this disk?

Any hardware errors being logged by OpenVMS?

Re: question on mount verification loop

Shriniketan Bhagwat — Wed, 28 Jul 2010 02:19:49 GMT

Hi,

A device enters into mount verification state when device becomes inoperative. This might occur in the following instances:

1. The device is accidentally placed off line.
2. The active port of the drive fails.
3. A hardware error occurs.
4. The device is set to write protected during a write operation.

When the device comes back on line, either through automatic failover or operator intervention, the operating system validates the volume.

Regards,
Ketan

Re: question on mount verification loop

Shriniketan Bhagwat — Wed, 28 Jul 2010 02:37:16 GMT

Hi,

A mount-verification is triggered by certain classes of disk IO errors. OpenVMS tries to re-establish or verify the IO path to and the contents (volume label etc.) of the disk, which caused the IO error. To further analyze these mount-verifications and trying to find the underlying reason, look at OPERATOR.LOG entries and ERRLOG entries (using DECevent or WEBES/SEA) on the OpenVMS systems.

Regards,
Ketan

Re: question on mount verification loop

Shriniketan Bhagwat — Wed, 28 Jul 2010 02:42:07 GMT

Hi,

Please refer the below link. It explains typical causes for mount verification.

http://h71000.www7.hp.com/wizard/wiz_9804.html

Hope this was helpful.

Regards,
Ketan

Re: question on mount verification loop

Shriniketan Bhagwat — Wed, 28 Jul 2010 03:03:13 GMT

Hi,

Few more links on similar topics.

http://h71000.www7.hp.com/wizard/wiz_8101.html
http://h71000.www7.hp.com/wizard/wiz_9095.html
http://h71000.www7.hp.com/wizard/wiz_8764.html

Also there are lots of discussions on similar topics in ITRC. Search them using the string mount verification.

Regards,
Ketan

Re: question on mount verification loop

Volker Halle — Wed, 28 Jul 2010 05:27:50 GMT

Antonio,

if you can boot from the OpenVMS I64 V8.3 CD, try mounting your 'failing' system disk.

MOUNT/FOREIGN will just access the disk. It will probably fail, if it can't read the disk. If it succeed, dismount the disk and try to fully MOUNT it.

$$$ MOUNT/FOR DKA100:
$$$ DISM/NOUNL DKA100:

$$$ MOUNT/OVER=ID/NOASSI DKA100:

Mount-verifications only happen on fully mounted disks as a reaction to certain disk-IO errors.

Volker.

Re: question on mount verification loop

Andy Bustamante — Wed, 28 Jul 2010 06:29:42 GMT

Make sure you don't have an error light on the disk. Power down and/or swap the disk to a different slot.

And of course try an updated release of VMS and make sure your patches up to date.

Did this disk log errors before failure? And finally, do you have a current backup of the disk?

Re: question on mount verification loop

Hoff — Wed, 28 Jul 2010 12:02:16 GMT

I don't think Ketan knows. :-)

Re: question on mount verification loop

Hoff — Wed, 28 Jul 2010 12:03:19 GMT

(about the origin of all that cited stuff, that is.)

Re: question on mount verification loop

antonio elder prado — Wed, 28 Jul 2010 14:53:19 GMT

Hi guys,

thank you very much for the info, i will be trying this afternoon and will submit the results later to the forum.
once again,

thank you very much, / muito obrigado,
./antonio/.

Re: question on mount verification loop

antonio elder prado — Wed, 28 Jul 2010 17:33:51 GMT

Hi,

i did check the machine and still got the mount verification messages.
i did boot out of the cd,
and under the prompt $$
did what Volker suggested, the only problem i see is on the last mount, it says something about bad label, and repair, then gives me the prompt back.
if i do a show device dka100/full it shows me mounted device.
another strangeness:
after long time of those messages, i can see in between the login, i do type the system and its password, and i can have the $ prompt (of course comes with those messages).
anyway,
i will keep on trying and if you have any extra suggestions it would be greatly appreciate,
thank you/obrigado.
./antonio/.

Re: question on mount verification loop

Andy Bustamante — Wed, 28 Jul 2010 17:58:17 GMT

Please try

$ mount ovms$dka100: /over=ident

from the $ prompt. If someone has modified either the startup file which mounts the disk or modified the volume lablel the automated startup will not be able to mount the device.

Re: question on mount verification loop

Andy Bustamante — Wed, 28 Jul 2010 18:39:31 GMT

>>>it says something about bad label, and repair, then gives me the prompt back

What is the message?

Re: question on mount verification loop

Hoff — Wed, 28 Jul 2010 18:55:17 GMT

V8.2 was _the_ first public release of OpenVMS for Itanium, and that release shipped out on January 13, 2005.

Accordingly, the default (shotgun) answer....

Upgrade to V8.3 or V8.3-1H1; to a newer and more current and likely more stable release.

And yes, this could well (still) be indicative of a hardware problem.

Heat, flaky disk, etc.

Which means the first part of the upgrade is a full and complete disk BACKUP of anything you need to save.

Given you have possibly-flaky storage or flaky system hardware, I'd start with a full backup of DKA100: and any other disks you have data on, regardless.

Re: question on mount verification loop

antonio elder prado — Wed, 28 Jul 2010 19:28:40 GMT

okay;

i spoke with one of the developers and we need to have the lowest possible OS available (in this case 8.2), so i need to stick with 8.2 crossing my fingers.

regarding the errors:
on the boot from the cd / 8.3
$$$ MOUNT/FOR DKA100:
the output of this mounting is:
MOUNT-I-MOUNTED, I64SYS mounted on _DKA100;

$$$ DISM/NOUNL DKA100
just got the $$$ back

$$$ MOUNT/OVER=ID/NOASSI DKA100
*** this time the error comes back as ***
MOUNT-W-IDX MAPPER
I/O ERROR ON INDEX FILE BITMAP, VOLUME LOCKED
SYSTEM-F-PARITY, PARITY ERROR
MOUNT-I-MOUNTED, I64SYS MOUNTED ON _DKA100;
$$$

thank you all for helping me with this,

./antonio/.

Re: question on mount verification loop

Andy Bustamante — Wed, 28 Jul 2010 20:40:33 GMT

The most likely out come is that anything on this disk has been lost.

You can attempt to load the latest ECOs for 8.2, get the current update plus any fiber_scsi and sys patches. This may or may not be productive in correcting this situation, but may help prevent the situation again. Don't expect this to resolve the issue.

Once you've decided data the disk isn't going to be recovered, you could attempt to initialize the disk and mount again. If this is hardware issue, the disk will soon fail again. Watch the error counts ($ show error) and check the error log for details. See the help for anal/error/elv for how to read error logs. If you start logging errors, replace the disk. If you have hardware support, replace it now.

Re: question on mount verification loop

P Muralidhar Kini — Thu, 29 Jul 2010 02:34:13 GMT

Hi Antonio,

>> MOUNT-W-IDXMAPPER
Here's the DCL help for IDXMAPERR

IDXMAPERR, I/O error on index file bitmap; volume locked

Facility: MOUNT, Mount Utility

Explanation: An I/O error was encountered during a mount operation.
Existing files in the volume can be accessed, but other
operations (for example, create and delete) are not possible.

User Action: Repair the volume with the ANALYZE/DISK_STRUCTURE/REPAIR
command.

But in your case the disk is getting in to mount verification during the boot
operation and hence you would not get a chance to run
ANALYZE/DISK_STRUCTURE/REPAIR on the disk.

>> SYSTEM-F-PARITY
Looks like there is some problem with the disk.
Did the disk work before ?
When did you start seeing the problem ?

Also as Andy had mentioned, you need to look out for the error count field in the
"SHOw DEVICE" output of the device. If this number keeps on increasing then
its a sign of problem.

For now, looks like ther is a problem with the disk $dka100.
If you have another system disk in the node then you can boot the node with the
other system disk. And once the node comes up, you can run
ANALYZE/DISK/LOCK on the $dka100 disk to check for errors.

Hope this helps.

Regards,
Murali

Re: question on mount verification loop

Shriniketan Bhagwat — Thu, 29 Jul 2010 02:51:54 GMT

Hi,

SYSTEM-F-PARITY message indicates some problem with the disk. The exact meaning of this message depends on the device. Use the DIAGNOSE command to examine the system error log for more information related to an error occurring on the device. On the disk drives one or more of the following conditions can cause this error:

1. Drive parity error
2. ECC hard error
3. Header compare error
4. Map parity error
5. Header CRC error
6. MASSBUS control parity error

Regards,
Ketan

Re: question on mount verification loop

Shriniketan Bhagwat — Thu, 29 Jul 2010 02:59:22 GMT

Hi,

Please refer the below link on PARITY errors.

http://h71000.www7.hp.com/wizard/wiz_6021.html
http://h71000.www7.hp.com/wizard/wiz_6418.html
http://h71000.www7.hp.com/wizard/wiz_9580.html

Regards,
Ketan