iSCSI initiator discovery problem
Dear Experts,
Greetings of the Day !!
I'm trying to setup as samba failover cluster in CentOS 5.3 x86 for testing purpose. For that I've setup a Openfiler SAN storage (used a PC with 1 P4 CPU of 2 GHz, 1 GB RAM, 80 GB IDE HDD), 2 CenOS nodes for failover cluster. The details are
Openfiler Storage - IP (172.19.27.70)
volumes - quorum (1 GB)
- sanvol1 (8 GB)
Node1 - IP (172.19.27.111)
Node2 - IP (172.19.27.222)
All the systems are in network and can ping and SSH each other. ISCSI-initiator-utils are installed on both cluster nodes. Further in this process, First, i've discovered the mapped LUNs on node1.
[root@clus-node1 ~]# iscsiadm -m discovery -t sendtargets -p 172.19.27.70
172.19.27.70:3260,1 iqn.2010-12.com.openfiler:quorum
172.19.27.70:3260,1 iqn.2010-12.com.openfiler:sanvol1
[root@clus-node1 ~]#
[root@clus-node1 ~]# fdisk -l
Disk /dev/sda: 6442 MB, 6442450944 bytes
255 heads, 63 sectors/track, 783 cylinders
Units = cylinders of 16065 * 512 = 8225280 bytes
Device Boot Start End Blocks Id System
/dev/sda1 * 1 13 104391 83 Linux
/dev/sda2 14 204 1534207+ 82 Linux swap / Solaris
/dev/sda3 205 783 4650817+ 83 Linux
Disk /dev/sdb: 1073 MB, 1073741824 bytes
34 heads, 61 sectors/track, 1011 cylinders
Units = cylinders of 2074 * 512 = 1061888 bytes
Disk /dev/sdb doesn't contain a valid partition table
Disk /dev/sdc: 8623 MB, 8623489024 bytes
64 heads, 32 sectors/track, 8224 cylinders
Units = cylinders of 2048 * 512 = 1048576 bytes
Disk /dev/sdc doesn't contain a valid partition table
[root@clus-node1 ~]#
Secondly, then i discovered the volumes on second node..
[root@clus-node2 ~]# iscsiadm -m discovery -t sendtargets -p 172.19.27.70 172.19.27.70:3260,1 iqn.2010-12.com.openfiler:sanvol1
172.19.27.70:3260,1 iqn.2010-12.com.openfiler:quorum
[root@clus-node2 ~]#
[root@clus-node2 ~]# fdisk -l
Disk /dev/sda: 7516 MB, 7516192768 bytes
255 heads, 63 sectors/track, 913 cylinders
Units = cylinders of 16065 * 512 = 8225280 bytes
Device Boot Start End Blocks Id System
/dev/sda1 * 1 13 104391 83 Linux
/dev/sda2 14 274 2096482+ 82 Linux swap / Solaris
/dev/sda3 275 913 5132767+ 83 Linux
Disk /dev/sdb: 8623 MB, 8623489024 bytes
64 heads, 32 sectors/track, 8224 cylinders
Units = cylinders of 2048 * 512 = 1048576 bytes
Disk /dev/sdb doesn't contain a valid partition table
Disk /dev/sdc: 1073 MB, 1073741824 bytes
34 heads, 61 sectors/track, 1011 cylinders
Units = cylinders of 2074 * 512 = 1061888 bytes
Disk /dev/sdc doesn't contain a valid partition table
[root@clus-node2 ~]#
The problem is when I discovered the LUNS using iscsiadm util, it discovered the quorum volume first then sanvol1 (which I wanted it should be), but in the same time on Node2, it discovered the sanvol1 first then quorum volume. Due to this it, the device names of these volumes differs
i.e.
quorum on node1 - /dev/sdb
sanvol1 on node1 - /dev/sdc
whereas
sanvol1 on node2 - /dev/sdb
quorum on node2 - /dev/sdc
As far as i know that the discovered device named should be same on both nodes. Is there any way to prioritize volume discovery of volumes (i.e. quorum should discovered first then sanvol1 on both nodes), inresult it will assign the same device names on both nodes and should remain persistent.
Kindly suggest.
Regards
Arunabh
Last edited by arunabh_biswas; 12-13-2010 at 07:08 AM.
|