remove disk drives - had issues with new driver

[John Jimenez]

I opened up a thread earlier. A new lun was using an old driver. I thought the rmsf command was the resolution. when I re-introduced the lun and ioscan and insf created a new drivers

crw-rw-r-- 1 oracle dba 188 0x0d5000 Apr 9 10:12 c13t5d0
crw-r----- 1 bin sys 188 0x0f5000 Apr 9 10:12 c15t5d0
crw-r----- 1 bin sys 188 0x115000 Apr 9 10:12 c17t5d0
crw-r----- 1 bin sys 188 0x135000 Apr 9 10:12 c19t5d0

but I still am getting errors with the dd command.

maxux2:[/dev/rdsk]
# dd if=/dev/rdsk/c13t5d0 of=/dev/null bs=1024 count=100
dd read error: I/O error
0+0 records in
0+0 records out

Did I need to reboot after issuing rmsf command to all 4 paths?

rmsf /dev/rdsk/c13t5d0
rmsh -H 1/0/2/1/0.97.4.239.0.5.0

[Steven E. Protter]

Shalom,

insf -C disk

Detects new disks.

When referring to an old thread, please provide a link so people can read it and catch up.

SEP

[John Jimenez]

Sorry SEP here is the previous thread

http://forums13.itrc.hp.com/service/forums/questionanswer.do?threadId=1331163

[Pete Randall]

John,

I have to wonder - in both your previous thread and this one, you mention a single lun. Is that true? Is there just one lun involved here? If that's the case, I suspect that you're not even looking at the new lun and need to re-examine your ioscan output to identify the new lun.


Pete

[John Jimenez]

When I issued the insf command for the 2 LUNS (disks) I was having issues with today, it did create new special files. Yesterday before doing the rmsf this did not.

I did not include -e on the insf because when I use the -e it changes the ownership of disks that are owned by oracle and changes them to bin and everyone gets locked out of the databases I wonder if -C will do the same thing.


maxux2:[/dev/rdsk]
# insf
insf: Installing special files for sdisk instance 202 address 1/0/2/1/0.97.4.239.0.5.0
insf: Installing special files for sdisk instance 205 address 1/0/2/1/0.97.4.239.0.5.1
insf: Installing special files for sdisk instance 151 address 1/0/2/1/0.97.8.239.0.5.0
insf: Installing special files for sdisk instance 152 address 1/0/2/1/0.97.8.239.0.5.1
insf: Installing special files for sdisk instance 201 address 1/0/4/1/0.97.4.239.0.5.0
insf: Installing special files for sdisk instance 203 address 1/0/4/1/0.97.4.239.0.5.1
insf: Installing special files for sdisk instance 153 address 1/0/4/1/0.97.8.239.0.5.0
insf: Installing special files for sdisk instance 155 address 1/0/4/1/0.97.8.239.0.5.1

[John Jimenez]

No actually I was having issues with 2 LUNS, I removed all 8 paths (4 to each LUN)

maxux2:[/dev/rdsk]
# rmsf /dev/rdsk/c13t5d0
maxux2:[/dev/rdsk]
# rmsf /dev/rdsk/c15t5d0
maxux2:[/dev/rdsk]
# rmsf /dev/rdsk/c17t5d0
maxux2:[/dev/rdsk]
# rmsf /dev/rdsk/c19t5d0
maxux2:[/dev/rdsk]
# rmsf /dev/rdsk/c13t5d1
maxux2:[/dev/rdsk]
# rmsf /dev/rdsk/c15t5d1
maxux2:[/dev/rdsk]
# rmsf /dev/rdsk/c17t5d1
maxux2:[/dev/rdsk]
# rmsf /dev/rdsk/c19t5d1
maxux2:[/dev/rdsk]
# rmsh -H 1/0/2/1/0.97.4.239.0.5.0
sh: rmsh: not found.
maxux2:[/dev/rdsk]
# rmsf -H 1/0/2/1/0.97.4.239.0.5.0
maxux2:[/dev/rdsk]
# rmsf -H 1/0/2/1/0.97.8.239.0.5.0
maxux2:[/dev/rdsk]
# rmsf -H 1/0/4/1/0.97.4.239.0.5.0
maxux2:[/dev/rdsk]
# rmsf -H 1/0/4/1/0.97.8.239.0.5.0
maxux2:[/dev/rdsk]
# rmsf -H 1/0/4/1/0.97.4.239.0.5.0
maxux2:[/dev/rdsk]
# rmsf -H 1/0/2/1/0.97.4.239.0.5.1
maxux2:[/dev/rdsk]
# rmsf -H 1/0/2/1/0.97.8.239.0.5.1
maxux2:[/dev/rdsk]
# rmsf -H 1/0/4/1/0.97.4.239.0.5.1
maxux2:[/dev/rdsk]
# rmsf -H 1/0/4/1/0.97.8.239.0.5.1
maxux2:[/dev/rdsk]

[John Jimenez]

I removed the disk from the host and am going to plan on removing the drivers again.

This time I will take SEP's suggestion and use insf -C

So will the command look like this

insf -C disk

[John Jimenez]

Doing the insf -C assigned new drivers, just like insf by itself did. But the dd command would still not run successfully.
I am figureing that maybe the systems need to be bounce to clear up something thats still there.

So, I finally wound up deleting the 2 LUNS that I created last week and re-creating the new LUN's wit a new number, in order to get new c#t#d# disk numbers. That did the trick.

I figure that I will probably get to use the other c#t#d#, after the system is bounced.

thanks again for you help.