TOPIC:

Thanks Bill & Salvatore,

After reboot my problem is solved. i have 3 iscsi on xen now.

I'm preparing to deploy an additional iSCSI in a two node Ha-Lizard scenario that I maintain for a customer. What I have is probably a stupid question afterall, but I was unable to reproduce it on a virtual environment.

All the examples that I could find in the threads always mention devices that are in the same addresses in both nodes (/dev/sda3 when it's the first disk, /dev/sdb for the second and so on). In this real scenario that I have, Server 1 has the new drive on /dev/sdb while Sever 2 has it on /dev/sde. I tried to switch bays and reboot this second node but it's somehow a little stubborn and it's never showing up as /dev/sdb, although it doesn't have anything occupying /dev/sdb (or even sdc and sdd).

Should I expect any issue with that?

I'll probably end up with a /etc/drbd.conf like this:

on xcpsrv01 {
device /dev/drbd2;
disk /dev/sdb;
address 10.10.10.1:7790;
meta-disk internal;
}
on xcpsrv02 {
device /dev/drbd2;
disk /dev/sde;
address 10.10.10.2:7790;
meta-disk internal;
}

Thank you for any inputs.

I found the answer in a sample file of another user in this same thread
(halizard.com/media/kunena/attachments/1949/second_iscsi.txt)

This will work fine, however, you will also need to update the LVM filter in 2 files to match the new device names in order to block LVM from reading LVM metadata directly from the block device.. LVM metadata should only be read via the iSCSI mount.

/etc/lvm/lvm.conf
/etc/lvm/master/lvm.conf

The only small side affect of this is that the iscsi-ha backup/restore scripts will not be reliable since they expect the exact same configuration on both nodes. If you are not using this feature, then you will not be losing any functionality you already have.