Intermittent I/O error on a HBA card.

[John Goetz]

Getting the following error in dmesg:
I/O hardware probe timed out
Last successful probe at path 0/6/1/0/4/0.97.40.19.0.0

The above path comes and goes. Ioscan, pvcreates, etc. takes a long time.

This is a 2GB SAN / 1GB Ethernet card. Model A9784A running on HP-UX 11.11.

[RAC_1]

Run STM and check the card. Also it would be useful to at lanadmin stats of the card.

Anil

[Florian Heigl (new acc)]

please check the card for transmission errors - get the device file from ioscan -fnkCfc and then issue /opt/fcms/bin/tdutil /dev/tdN stat - take the full output and save it.
afterwards reset the error counters and generate I/O for approx 1h - compare it - did the Bad RX chars or sync loss counters increase?

also don't just (loopback) test the card from STM but also log into 'logtool' inside STM, where close to any lowlevel fc errors get logged.

I'd say You should even find a lot of errors in dmesg!

[Florian Heigl (new acc)]

hehe - just saw You got the error from dmesg - I'm too tired it seems.


next steps would be replacing the switch port, switch port gbic cable, opening a call with hp, off- and onlining the port and at worst having the adapter replaced.

out of curiousity - is it a tachyon XL2?