TOPIC:

Dear friends,
I have a problem with my HA structure.

Scenario
- 2 servers running HA-Lizard and sharing an iSCSI SR
- First server, xenserver1, it's running 4 VM's. It's the master of the HA and of the iSCSI-HA.
- Second server, xenserver2, it's running 2 VM's. And it's the slave of the HA and iSCSI-HA.

At first sight, the iSCSI looks good at the Xencenter.
My 4+2 VM's are running fine because everything goes well: good power supply, no hardware issues, ...
The goal to have this structure is to have all the VM's running on one server in case that the other one fails or run into maintenance mode.

In the past I've moved VM's between servers without problems. From xencenter1 to xencenter2 and viceversa.
But few days ago I've tried to move one of the 4 VM's running on xencenter1 to xencenter2 just to get a load balancing structure.

I've tried to migrate my "less important" VM: losvpxe. I've not tried with the other ones.

Doing that migration using xencenter (Xencenter > migrate to server > xenserver2) it looks like the machine migrates normally. But if you login at the VM, after few seconds of few commands the machine it's hanged.
If I shutdown the VM on xenserver1 and if I start it on xenserver2, you can read that there's no hard disk to boot for that machine.

I don't know if that problem happens only with that VM or if all of them are affected.
I don't know if the problem only happens if I try to migrate from xenserver1 to xenserver2... or if in the reverse also happens

Please, see attached pictures for more information.

If there is any extra information that I can provide to help to solve the problem do not hesitate to ask me.

Thanks in advance.

Everything in your screenshots looks normal. Any chance you are running xcp-ng 7.6? There is a known bug in that version which breaks live migration.

Thanks Salvatore for your reply.
I don't know how to check what you say about xcp-ng 7.6. How can I check it?
In the past I have used that live migration feature. But in any case I also have tried to shutdown the VM from xenserver1 and run it on xenserver2, with the "no disk to boot" message. That's not a live migration I think.

xcp-ng is a rather new alternative to xenserver. Can you confirm which version of XenServer you are running?

Incorrect LVM filters could cause this type of issue.
Can you post the "filter" line from the following files on both hosts.
/etc/lvm/lvm.conf and /etc/lvm/master/lvm.conf

Is that correct?

---

[root@xenserver1 ~]# more /etc/lvm/lvm.conf | grep filter | grep -v "#"
filter = [ "r|/dev/sda3|", "r|/dev/drbd1|" ]

---

[root@xenserver1 ~]# more /etc/lvm/master/lvm.conf | grep filter | grep -v "#"
filter = [ "r|/dev/sda3|", "r|/dev/drbd1|" ]

---

[root@xenserver2 ~]# more /etc/lvm/lvm.conf | grep filter | grep -v "#"
filter = [ "r|/dev/sda3|", "r|/dev/drbd1|" ]

---

[root@xenserver2 ~]# more /etc/lvm/master/lvm.conf | grep filter | grep -v "#"
filter = [ "r|/dev/sda3|", "r|/dev/drbd1|" ]

---

Thank you.

The filters are correct. Can you also confirm that the VM you are migrating has its storage on the iscsi SR (i see there is also an NFS SR in your pool).