I inherited an IBM System Storage DS4000 SAN with an AIX environment running Oracle RAC (3 nodes). One of the nodes crashed, so I got the system up and running with 2 nodes; however, the node that crashed was the node that was managing the backups. The backups were stored on a LUN in the SAN.
Due to the lack of documentation, I can see how they're mapped from
/etc/filesystems, but I'm not sure how the AIX environment mapped these LUNs at bootup (other than the mapping association from the SAN management interface). I want to re-map the LUNs used for backups from the old node to one of the remaining nodes, but I'm concerned about losing the data on the LUNs.
I found some instructions here, but that seems like it's only for new LUNs. If I re-map the LUN to a new host, would the existing data on that LUN stay intact?
It looks like I knew less than I thought.
I created a new array on my storage controller called test and a new LUN called logbackup to test my theory and mapped it to one of the nodes as LUN 12.
On that node, when I ran
cfgmgr, I ran into an error and followed these instructions to resolve it.
With cfgmgr working, I thought I'd be able to see the LUN, but I don't--or at least, the output of
lsdev -Cc Disk don't give me the information I can easily translate.
I've unmapped it and remapped the LUN a few times, but there is no change to the list of hdisks when I run cfgmgr.
Looks like I was reading my
lscfg output incorrectly. I changed the LUN from 12 to 255 and saw a change in the
lscfg. It mapped the LUN to hdisk16 so I added it as a physical volume using
chdev -l hdisk16 -a pv=yes. Then, running
fget_config -Av, I was able to confirm the LUN mapped properly. Now, it's just a matter of mounting it, writing to it, then migrating it to my other node to see if the data stays intact.