How to reset device error count on OpenVMS V6.2-1H3

[Gordon Morrison_1]

I found this command in another thread, but it doesn't work:

$ SET DEVICE/RESET=ERROR_COUNT MKA100:
%DCL-W-IVQUAL, unrecognized qualifier - check validity, spelling, and placement
\RESET\
$

I assume this is because it's such an old version. Is there an equivalent command in V6.2 ?
Thanks

[labadie_1]

Hello

Zdec works with VMS 6.2

It is listed
http://h71000.www7.hp.com/openvms/freeware/v8_00freeware_abstract.txt

It should be in the disk3 of the VMS Freeware
http://h71000.www7.hp.com/openvms/freeware/

Regards

[labadie_1]

It is available at
http://www.tmk.com/ftp/vms-freeware/fileserv/

[labadie_1]

and clear_errors

http://vms.process.com/scripts/fileserv/fileserv.com?CLEAR_ERRORS

Keep in mind that this stuff means a possibility of crashing your machine.

[Gordon Morrison_1]

Thanks, but this is a sensitive system and changes like installing new software would never get approved.
How can I find out if it's already installed, and what would be the command to reset the error count?
(Sorry, Unix admin with limited VMS)

[Gordon Morrison_1]

"Keep in mind that this stuff means a possibility of crashing your machine."

:-O Definitely not an option.

[labadie_1]

if this is not an option, plan a reboot of your computer.

To the best of my knowledge, there is not an HP approved tool existing for VMS 6.2-1H3, and I think there will never be.

[Gordon Morrison_1]

OK, Thanks.

[Gordon Morrison_1]

Question answered.

[Andy Bustamante]

Gordon,

The other option of course is to simply note the current error count and watch for changes. There's no clear management reason that this should be 0.

Andy

[Jon Pinkley]

Gordon,

The supported way to clear device error counts in VMS 6.2 is to reboot. As others have said, the change in value is more interesting than absolute error count.

"set device/reset" actually writes an errlog message recording the fact that it reset the device error count. As far as I know, none of the user-contributed error resetting programs do this (and it they did, they probably did it in a different way).

Any software that runs with cmkrnl or cmexec privilege enabled has permission to crash a system. I have used ZDEC and looked at the code and from what I remember, it was written with safeguards in place to prevent crashing the system.

You can also use the kernel debugger DELTA to clear the error counts, and this requires no "extra" software to be installed, but doing that is in my opinion much more dangerous than using a utility like ZDEC.

The safest is to just live with the error counts until you reboot.

Since your example used a tape device, what is your goal in resetting the error count?

Are you trying to determine if the tape drive needs cleaned? If so, depending on the drive there may be better ways to get the info (using SCSI commands passed directly to the tape drive via the MKDRIVER SCSI pass-through).

In our backup command procedures, we get the device error count before backup starts, and again after the backup is complete. The differences are what we want to be aware of. A bad tape, or poor SCSI cable connection can cause a lot of tape errors, but as long as the absolute number does not increase during the backup, then we don't worry about it.

Jon