There are times at certain environment that the supposed to be fixed IP Address
of the shared storage have to change.
Althought ,am having difficulties to contemplate the reality that it actually changed twice!

so, from the earlier on, the factory setting is something like 192.168.0.1/24
then goes to another set of “fixed” IP Address.
And now the third round of this..

the result, whenever this one changed.. all the client trying to attach to this
particular iSCSI storage, it need to probe for new IP.
And as past IP setting .. it was saying “portal unreachable”..
in this case only one out of two setting is up.
resulting a failure in status.. the storage not mounted..

[root@node1~]# /etc/init.d/iscsi restart
Stopping iSCSI daemon: 
iscsid dead but pid file exists                            [  OK  ]
Starting iSCSI daemon:                                     [  OK  ]
                                                           [  OK  ]
Setting up iSCSI targets: Logging in to [iface: default, target: iqn.2004-08.jp.buffalo:LOMBOK-DB-001D7326325B:array1, portal: 192.168.1.145,3260]
Login to [iface: default, target: iqn.2004-08.jp.buffalo:LOMBOK-DB-001D7326325B:array1, portal: 192.168.1.145,3260]: successful
 Logging in to [iface: default, target: iqn.2004-08.jp.buffalo:LOMBOK-DB-001D7326325B:array1, portal: 192.198.1.148,3260]
Login to [iface: default, target: iqn.2004-08.jp.buffalo:LOMBOK-DB-001D7326325B:array1, portal: 192.198.1.148,3260]: failure
                                                           [  FAILED  ]

so, the question.. how to remove those invalid target ?
Here is how to delete old target.
replace the ip with the old IP setting.

iscsiadm -m node -p 192.198.1.148 --op=delete

to add the new one ..

iscsiadm -m discovery -t sendtargets -p 192.168.1.145

after that tweak the udev rules ..

vi /etc/udev/rules.d/buffalo.rules

type the following content to it..

  KERNEL=="sd*", SUBSYSTEMS=="scsi", ATTRS{model}=="VIRTUAL-DISK", SYMLINK+="usbhd%n

so.. it would sym-link the iSCSI drive to

/dev/usbhd1
udevcontrol reload_rules
# restart iscsi..
/etc/init.d/iscsi restart 

see the new dev ..

shall had something like this..

 
ll /dev
lrwxrwxrwx 1 root root          3 Dec 22 23:02 usbhd -> sdb
lrwxrwxrwx 1 root root          4 Dec 22 23:02 usbhd1 -> sdb1

which later can be used by cluster suite to mount /dev/usbhd1 to something like /data
which is transparent/similar on both nodes regardless it had the USB drive/dongle or anything connected to its
system at the same time.

Related posts:

  1. Rsync over SSH for restoring backup purpose
  2. How to compile custom kernel for Vortex86 SX
  3. System-config-network crashed on CentOS 5.3
  4. How to connect GPRS via Bluetooth on k530i in CentOS
  5. Google Chrome OS under the hood screenshot