OLA/R commands, auto resumed after replacement?

[bennethos]

Hi all,

I have a question concerning a problem I had with replacing a faulty FC card using OLA/R commands.

Putting on the attention LED and powering off the slot + suspending the driver worked like a charm... I went to the server and the light of the pci slot was blinking, there was no single activity/LED on the card that had to be replaced.

I removed the card, put in the new card and it started working immediately,leds went on, TX was emitting light etc...

This shouldn't have happened, nobody resumed the card. here's my logfile

So basically my question is, why did hp-ux did a poweron/resume of the card automatically ?

Oct 18 14:06:09 s0022640 vmunix: 0/0/8/1/0/4/0: Fibre Channel Driver
Received Suspend Request.
Oct 18 14:06:09 s0022640 vmunix: 0/0/8/1/0/4/0: Fibre Channel Driver
has been Successfully Suspended.
Oct 18 14:05:29 s0022640 vmunix: 0/0/8/1/0/4/0: Fibre Channel Driver
at /dev/fcd9 is being reset
Oct 18 14:06:09 s0022640 vmunix:
Oct 18 14:20:12 s0022640 above message repeats 3 times
Oct 18 15:19:49 s0022640 vmunix: 0/0/8/1/0/4/0: Fibre Channel Driver
Received Resume Request.
Oct 18 15:19:49 s0022640 vmunix:
Oct 18 15:19:49 s0022640 vmunix: 0/0/8/1/0/4/0: Fibre Channel Driver
has been Resumed Successfully.
Oct 18 15:19:50 s0022640 vmunix: 0/0/8/1/0/4/0: Fibre channel host
port is OFFLINE, can not scan
Oct 18 15:19:50 s0022640 vmunix:
Oct 18 15:20:12 s0022640 above message repeats 2 times
Oct 18 15:23:58 s0022640 vmunix: 0/0/8/1/0/4/0: Fibre Channel Driver
received Link Dead Notification.
Oct 18 15:23:58 s0022640 vmunix:

I used the following commands :

# rad -q
# rad -f ATTENTION 0-0-0-2
# rad -o 0-0-0-2
# rad -s 0-0-0-2 suspend

That's all I did...

thx

[Torsten.]

>> This shouldn't have happened, nobody resumed the card. here's my logfile


This depends. Most servers have a switch, door lock or whatever for each slot.

Open this switch does the same like a power off command, closing gives power.

[bennethos]

The system was an rp7420

[Torsten.]

I assume you pressed the attention button while closing the latch.

[Torsten.]

Anything in this file?

/var/adm/hotplugd.log

[bennethos]

Yeah there's indeed a button with a BELL symbol on it... it's possible I pressed it accidentally ;-) , i'll check the logfile

[bennethos]

output of log :

hotplugd : Oct 18 15:19:44 2009 MET : #1 slot(0-0-0-1) - oprn(POST_REPLACE)
- olarIoEvent(INSERTION_UNPOWERED) - flags() Attention Button Event
Received -
rcvd-time(Oct 18 15:19:44 2009 MET)
hotplugd : Oct 18 15:19:44 2009 MET : #1 POST_REPLACE - slot(0-0-0-1) -
time(Oct 18 15:19:44 2009 MET)
Activity : Start of Post Replace
Target slot : 0-0-0-1

Activity : post_replace:/usr/sbin/olrad.d/igelan driver script not
present
Target slot : 0-0-0-1
Hardware path : 0/0/8/1/0/6/0

Activity : End of Post Replace
Target slot : 0-0-0-1

Activity : Target slot powered on, drivers resumed, OK to start using
the card
Target slot : 0-0-0-1

hotplugd : Oct 18 15:19:51 2009 MET : #1 POST_REPLACE - slot(0-0-0-1) -
Completed
hotplugd : Oct 18 15:19:51 2009 MET : Waiting for PCI I/O Attention Button
event...

[Torsten.]

Looks like you pressed the button:


Attention Button Event Received

[bennethos]

so installing the card and pressing the button makes the card go poweron and the driver state resumed ?

What would have happened if I had press a button of an active card I wonder ?

thx for the info so far!

[Torsten.]

This will deactivate and power off the card.

In the beginning of OLAR the server (rp54xx and rp7400) doesn't have such buttons, so you need to perform the commands. All later models have such buttons and switches to make the replacement easier (see appropriate user service manuals).