Control number (device file), Ignite restore and MC ServiceGuard

[Nikee Reddy]

Hello Folks,

I have a question related to the device file (control number)and the MC ServiceGuard configuration.

Normally the device file number changes after the Ignite restore. Since the c? is the hardware-bus id.

In a non-cluster environment I can the device file number by using the following commands.

#ioscan -f | grep -e INTERFACE | grep -v target | awk '{print $3, $1, $2 }' > infile

#sbin/ioinit -f /infile -r
----reboot----
#insf -e

Someone told me that the above commands/procedure is not recommended in the cluster environment. Is there any other better procedure for this to do?

I really appreciate your response.


Thanks,
Nikee

[Michael Steele_2]

Your friend must be refering to :

1) Matching the same devices on both nodes, and,

2) Re-making the cluster.ascii file to accomodate the changes.

The cluster binaries control all the resources with MC/SG so they would have to be remade.

There is one more possibility, your procedure is incomplete.

A) #ioscan -f | grep -e INTERFACE -e DEVICE | grep -v target | awk '{print $3, $1, $2 }' > /infile

(* Note: infile resides in / *)

B) # vi /infile and modify

# shutdown -r y

# ISL> hpus -is (* single user mode *)

# /sbin/ioinit -f /infile -r
# auto reboot takes place

###########################################
# TWO
###########################################

1) A) above
2) B) above
3) mv /stand/ioconfig /stand/ioconfig.sav
mv /etc/ioconfig /etc/ioconfig.sav
shutdown -r now
4) Do to missing 'ioconfig' file the server will stop at: (IOINTRC:>
5) (IOINTRC:> # /sbin/ioint -c
6) (IOINTRC:> # /sbin/ioinit -f /infile -r
7) auto reboot takes place

Make sure you have two make_tape_recovery tapes verified before doing this.