Aditional information from Udo de Boer:
"I would like to comment you on your summary. All compaq external raid
controllers
like the HSZ70, HSZ80, HSG80, ect report their errors to the host system.
These
errors will come up in the messages file but also in the binary errorlog.
The binary
error log you can view with "dia -R |more" . Dia is part of decevent and
decevent is
located on the asocciated product cd's. You can also download it at compaq.
The
latest version of dia I think is 3.3. It could be that internal raid
controllers are
also supported.
If you want to have more control are want to have email messages use the
Storage
Works Command Console. This is a software agent that runs on the system and
can also
send mail during errors. There is also a gui included that runs on Windows
that
allows you to see the raid controller. This should work for all compaq raid
controllers. StorageWorks Command Console is delivered together with the
raid
hardware and newer versions can be downloaded."
Thanks
-----Mensaje original-----
De: tru64-unix-managers-owner_at_ornl.gov
[mailto:tru64-unix-managers-owner_at_ornl.gov]En nombre de jgrijalba
Enviado el: martes, 10 de julio de 2001 9:50
Para: Tru64-Unix-Managers (E-mail)
Asunto: SUMMARY: RAID disk check
Hi gurus,
Original question:
How to check a RAID disks looking for errors.
Answers:
Thanks to Alan, John Tan, Kjell Andresen and Thomas Blinn for their
answers.
I choose the answers from Thomas Blinn and Alan:
Thomas :
"Read the RAID controller documentation. With most RAID controllers, the
host system (e.g., Tru64 UNIX) can NOT see the "real" disks and never is
informed of most disk errors; you have to ask the RAID controller, and
the details of how to do that vary"
Alan :
" Some RAID systems may include their own software, so check
the subsystem documentation. If the RAID pretends to present
SCSI disks back to the system you can use scu(8) to run a
verify on the device. Read the help for the particular
command carefully. One of the scu(8) scanning commands
does a write followed by read, while the other just reads.
Using the wrong one, will overwrite all the data on the
disk. For non-SCSI disks, you can use the dd(1) on the
C partition of the disk. If the RAID has regenerate any
data, you'll have to consult the subsystem's documentation
to see if logs that somewhere."
Regards,
Jorge Grijalba
Received on Wed Jul 11 2001 - 08:02:12 NZST