Hi,
I have Suse Linux Enterprise Server 10 installed on two node ocfs2 cluster my server. Node-1 is ptdmod01 and Node-2 is ptdmod02. It is having 218 ocfs file system, which are shared among both servers. Since this is ocfs2, each disk from EMC Storage is mounted individually on file-system as below
PHP Code:
root@ptdmod01:~> df -h | tail -10
/dev/emcpowerhj 18G 246M 18G 2% /pcps/apps/test
/dev/emcpowerhk 18G 230M 18G 2% /vror/apps/test
/dev/emcpowerhl 8.8G 140M 8.6G 2% /inf/duet_dom/inf_domain02/cache/VROR/test
/dev/emcpowerhm 8.8G 148M 8.6G 2% /inf/duet_dom/inf_domain02/logs/VROR/test
/dev/emcpowerhn 8.8G 136M 8.6G 2% /inf/duet_dom/inf_domain02/cache/PCPS/test
/dev/emcpowerho 8.8G 139M 8.6G 2% /inf/duet_dom/inf_domain02/logs/PCPS/test
/dev/emcpowerhp 18G 461M 17G 3% /redwood/apps/rwtt_tlinft
/dev/mapper/vg_local_redwood-lv_reedwood_apps_rwtt
8.0G 744M 7.3G 10% /redwood/apps/rwtt
/dev/emcpowerhr 18G 379M 18G 3% /inf/duet_dom/inf_domain02/cache/ndw/uat
root@ptdmod01:~> cat /etc/fstab | tail -10
/dev/emcpowerhh /pfcst/apps/qa ocfs2 _netdev 0 0
dmonpod13:/root_vdm_5/APPS01/SASQA /pfcst/etl-qa nfs defaults 0 0
/dev/emcpowergo /redwood/apps/rw8d ocfs2 _netdev 0 0
/dev/emcpowerhj /pcps/apps/test ocfs2 _netdev 0 0
/dev/emcpowerhk /vror/apps/test ocfs2 _netdev 0 0
/dev/emcpowerhl /inf/duet_dom/inf_domain02/cache/VROR/test ocfs2 _netdev 0 0
/dev/emcpowerhm /inf/duet_dom/inf_domain02/logs/VROR/test ocfs2 _netdev 0 0
/dev/emcpowerhn /inf/duet_dom/inf_domain02/cache/PCPS/test ocfs2 _netdev 0 0
/dev/emcpowerho /inf/duet_dom/inf_domain02/logs/PCPS/test ocfs2 _netdev 0 0
/dev/emcpowerhp /redwood/apps/rwtt_tlinft ocfs2 _netdev 0 0
Apart from it there are few vg also created on server, which are also shared among both ocfs cluster servers. Now, I have a challenge here. Our Storage is going to be migrated from DMX to vmax, so all the data has to be migrated to new storage and those new luns to be presented to server. I know, vg can be exported and imported and server will be able to see new luns and will mount them as expected. But when storage team will migrate data to new storage and present new luns, those will come in different emc sudo format. OS takes help from /etc/fstab to mount them, but after migration, emc device name will be changed to something else and mount will fail.
One way is to check each new lun and change entry in /etc/fstab, but it is a tedious job since there are large number of devices. Is there any way, if I can keep same.
For example, currently /dev/emcpowerdj is mounted on /inf/domain02/ora_dev. lun id is 37AD for this. On Storage side, they will copy data to new lun and will present that lun to server. I will be able to see that lun in inq output. But I expect its naming convention to be /dev/emcpowerdj, so that I do not need to edit /etc/fstab.
Hope, I am enough clear in explaining my problem.