Message on console - SAN connected Alpha with Tru64UNIX v5.1A

[A.W.R]

Hi,

I have an Alpha system connected to a SAN running Tru64UNIX v5.1A. From time to time I get the following messages appearing on the console:-

Aug 30 15:02:53 ALPHA1 vmunix: ccfg_MakeDeviceIdentWWID: Invalid device ID: 0x500308c001b3b020...
Aug 30 15:02:53 ALPHA1 vmunix: ccfg_MakeDeviceIdentWWID: Invalid device ID: 0x500308c001b3b026...
Aug 30 15:02:53 ALPHA1 vmunix: ccfg_MakeDeviceIdentWWID: Invalid device ID: 0x500308c001b3b02c...


I would really appreciate it if anyone can shed any light on why they are happening, what is causing them and what I can do to fix them?

Thanks
Andrew

[Rob Leadbeater · Accepted solution]

Hi Andrew,

What sort of SAN are you connected to ?

According to the device ID, I'm guessing that you've may have something ADIC based... I wonder whether you need to configure something specific for Tru64 on that.

Hope this helps,

Regards,

Rob

[Vladimir Fabecic]

First we need more details.
What storage you have?
OS patch kit level?
Any recent changes to SAN environment?

[Hein van den Heuvel]

>> From time to time

So this is while Tru64 is otherwise happily running? Or is is duing a boot or something?

My impression is that the San in announcing availability of a device not supported by Tru64, or not supported by your kernel.

Is the system 'missing' some storage connection or is everything there that should be there?

I see 3 possible directions of attack

1) Tell you storage folks (you?) to ZONE the san, or use selective presentation to make sure the Tru64 HBA's only get to see the devices meant for it.

2) Tell the storage guys you are suspecting stray device announcements and let them escalate the problem. Something might be just wrong.

3) See if you can teach Tru64 to accpet and ignore the device? Maybe if you booted GENVMUNIX and ran DOCONFIG at a time you suspect that those announcements recently happened that a driver can be found for it? (I doubt it though). Maybe someone else is smart enough to teach Tru64 to map that ID to something innocuous?

Or... just ignore it, only reporting back to the san folks once per month how many of those messages you received.

Hope this helps some,
Hein van den Heuvel (at gmail dot com)
HvdH Performance Consulting

[A.W.R]

Hi,

Thank you for all your input. I think Rob is right on the button. There is an ADIC library on the SAN with 3 tape drives.

Can I ask how you know it is an ADIC device?

THanks
Andrew

[Rob Leadbeater]

Hi Andrew,

Quite easy really, as the first part of a WWID is relate to the manufacturer.

So stick 500308c0 into Google and it throws up a support document on www.adic.com

Cheers,

Rob

[Hein van den Heuvel]

So now the question still stands... is the Tru64 system using that device, or supposed to be using it, or does it happen to live on the same san for use by others?

Hein.

[A.W.R]

Thanks for all your responses. I need to do some more investigation about how these devices are being used.

Regards
Andrew

[A.W.R]

I need to do further inveatigation as to how these devices are being used.

[DCBrown]

There was a bug in V5.1A prior to bl23 that would erroneously generate this message. If running a version prior to this patch kit level then upgrading to the latest patch kit will probably solve the problem.

The patch summary states: The patch fixes a problem and prevents 'ccfg_MakeDeviceIdentWWID:
Invalid device ID' messages from being generated when they should not be.


The bug happens when there are duplicate identifiers within the list returned by the device. (Inquiry page 83h). Review the wwid page data:

# scu
scu> sbtl {of device}
scu> show inquiry page wwid

If there are duplicate entries, then this is likely the problem. If not, then posting the output will make it easier to venture a guess as to what is happening.

Invalid device ID: 0x500308c001b3b020...

The WWID conforms to IEEE conventions. See http://standards.ieee.org/regauth/oui/tutorials/fibreformat.html (table 6). So in this case, the ident is NAA format 5h which means the next 6 nibbles are the OUI (organizationally unique identifier). Plugging 00308C into page http://standards.ieee.org/regauth/oui/index.shtml results in the entry

00-30-8C (hex) ADVANCED DIGITAL INFORMATION


Bud