Controler problem, or what?

[Rick Garland]

A quick glance through your attachment and I do not find a dev entry that would point to a specific device.

If you have stm loaded, you can browse through the various devices on the system and get a good indication of what is going on.

[Patrick Wallek]

I agree that there is not an entry in the attached log for any specific device. It does say BUS 2 though. Could this be a problem with the SCSI Card that is Bus 2? Without an entry for a specific device, this is difficultto figure out.

You could try increasing the timeout of all of the drives on your system and see if that helps. You can do this with the command pvchange. pvchange -t io-timeout-value /dev/dsk/c#t#d# I have heard it recommended that you change the io-timeout-value to something like 90. That value is in seconds.

[Patrick Wallek]

Another thought - Are those errors an isolated incident or does it happen on a regular basis?

[Patrick Wessel]

Marco,
Another point is the firmware of the ST34573WC disk. Ensure that the drives have at least HP11 (better is HP14). A ST34573WC with HP05 firmware can be responsible for bus hangs

[Jerry Lindsey]

Had a similar problem about 9 months ago on a K380/10.20. Replaced all cables, HSC interfaces, HSC expander, both controllers in the array, firmware was the latest. Still was logging errors. Transferred all the drives, the controllers and memory from the array chassis to a new chassis, and the problem went away. Hope you don't have to go to that similar extreme.

[Marco Hernandez_1]

Patrick Wallek,

Responding to your question,

These errors are very often, probably once a month, they are not isolated errors.

[Marco Hernandez_1]

Patrick,

By the way, the hard drives are firmware 14

[Patrick Wessel]

Marco,
You get just one SCSI reset a month? I wouldn't spend one more thought on it. Once a month is nothing critical!

[Patrick Wallek]

With just one error a month I think I agree with Patrick Wessel. You might try changing the SCSI timeout value on your disks. I have heard from other folks that this does help the SCSI Lbolt timeout.

[Marco Hernandez_1]

Patricks,

Do you think that replacing a SCSI hardrive every month, is very common?, I do not have the same problem in another machine that similar.

[Stefan Farrelly]

Replcing a hard drive every month is definitely not common. Each new disk should last you around 4-5 years of constant use before it starts to die (powerfail messages or scsi resets caused by it). You have 4 disks on your internal controller, and vg00 is there (I guess -- and vg00 is always very busy) so you are probably overloading the scsi controller, hence the occassional resets. increasing the timeout should sort you out, or move some of the disks of 10/0 to another controller.

[Patrick Wessel]

Marco,
The typical lifetime of a disk should be at least 3 years.
I thought you had a SCSI reset once a month. Sorry for that?

[Victor BERRIDGE]

Is it always the same drive (topoloy) that you are changing? on one machine (H50) I had It was the 3rd disk on the right of the bottom draw, every 10 months about, on another it was the lan on the mother board 3-4 times in less than 6 months, this one was great because since I moved the machine (3 meters from its old location) it worked well ever since and all this doesnt happen just with HP (I have some identical case with SUN, DEC)...
Do you believe in gremlins?

Good luck
Victor

[Patrick Wallek]

I also was replying with the understand that it was one SCSI reset per month. A SCSI reset does not necessarily indicate a bad disk. If you are replacing one disk per month, then that is extreme. If it is one SCSI reset error per month, then I wouldn't worry quite as much.