vmunix: ERROR

[Stijn V]

Hi all,

Since a while we do receive the following error in /var/adm/syslog/syslog.log on one of our servers. Error occurs every Saturday at 5AM since 5 weeks?

vmunix: ERROR: can't get hardware address for device.

The only link that I can make is that on the days that we do receive this error, we do have an issue with a backup device (this server is the DP cell manager), although the time doesn't match between both? And up to HP there is nothing wrong with the tape drive (tape library MSL6000)?

[Steven E. Protter · Accepted solution]

Shalom,

I kind of disagree with HP's assessment.

To get full access to the tape library, the stape driver is probably compiled into the kernel.

Every time the device goes offline you get the error.

It's kind of a Homer Simpson Doh issue now.

The device should not be going off line.

If the device is performing backups and restored properly the message can be ignored, but I'd push HP to make it go away.

Eventually, some part of the system, the fiber card, scsi cable or tape library is going to fail and need replacement, thats my prediction.

Perhaps start with a mstm, cstm or xstm hardware test.

SEP

[Stijn V]

I alreday logged a HW call for the server now.

But how can I make the link between this error message and the device (which device? the message doesn't specify it)?

Thanks.

[Steven E. Protter]

Turn the device off.

tail -f /var/adm/syslog/syslog.log

See if you can create a causal relationship to the error.

Then tell HP to comefix it. I think its the tape lib.

SEP

[Stijn V]

FYI,

Apparently, this vmunix error is generated during the creation of an Ignite image of the server itself... HP will investigate if we need some LVM patches.

So this means that this issue is not related to the backup issue. I was able to reproduce the backup error, apparently it is due to an Ignite image from another server. I think that when you launch an ignite, it scan all devices an for some reason it does something with the SAN tape drive resulting in a backup failure (Cannot write to device ([5] I/O error) on the server above.

Thanks for the help.

[Stijn V]

HP sais that the HBA is overloaded at that moment (when creating Ignite adn backup over the same HBA)....

[Stijn V]

Closed