TOPIC:

Hi to all.

Please if you could give me a basic road map in order to upgrade xenserver 6.2 to 6.5 and Halizard attach to them.

We have HA-Lizard version 1.7.8 and Version: IHA_1.4.5 of iscsci.

When upgrade Xenserver the drbd kernel change Rights? and now xen server 6.5 already its attached to the kernel modules?

Its the upgrade Version Xenserver mayor releases possible and reinstall HA-lizard without problem?? or i better stay with 6.2?

Thanks!!

*** Updated - added clearing of LVM cache to step 8 ***

Hi - it is possible to perform an upgrade from VS 6.2 -> 6.5 with no VM downtime (with HA-Lizard and iSCSI-HA). We have done this successfully and documented the exact procedure below. Once on 6.5 - DRBD kernel modules are no longer required. Only the userland tools for DRBD are required.

Detailed steps below:
(1) Migrate ALL vms from Master to Slave host

(2) Disable High availability
ha-cfg status

(3) Move the storage from the Master to the Slave host
iscsi-cfg manual-mode-enable (on both hosts)
iscsi-cfg become-secondary (on master)
iscsi-cfg become-primary (on slave)


(4) Insert XenServer 6.5 installation ISO in Master and reboot, then follow instructions to upgrade. After upgrade is complete - disable openvswitch and reboot
xe-switch-network-backend bridge
reboot

(5) After upgrade of Master is complete - Place installation tarballs for latest ha-lizard and iscsi-ha in /tmp - unpack
cd /tmp
tar -zxvf ha-lizard-*.tgz
tar -zxvf iscsi-ha-*.tgz

(6) Install iscsi target framework
yum --enablerepo=base --disablerepo=citrix install scsi-target-utils

(6) Install required packages
/tmp/ha-lizard-*/scripts/install --nostart
rpm -ivh /tmp/iscsi-ha-1.5.4/RPM/3.10.0+2/drbd-utils-8.4.3-2.x86_64.rpm
chkconfig drbd off
/tmp/iscsi-ha-*/scripts/install --nostart


(7) Copy config files from peer to upgraded host

scp /etc/lvm/lvm.conf root@<upgraded_host_ip>:/etc/lvm/
scp /etc/drbd.conf root@<upgraded_host_ip>:/etc/
scp /etc/tgt/targets.conf root@<upgraded_host_ip>:/etc/tgt/
scp /etc/iscsi-ha/iscsi-ha.conf root@<upgraded_host_ip>:/etc/iscsi-ha/
scp /etc/sysconfig/iptables root@<upgraded_host_ip>:/etc/sysconfig/
service iptables restart

( re-establigh replication on upgraded host
service drbd start
rm –f /etc/lvm/cache/.cache && vgscan

(9) force host into manual mode before starting iscsi-ha
> /etc/iscsi-ha/state/manual

(10) Start iscsi-ha - check status - should be in manual mode and secondary
service iscsi-ha start -w
iscsi-cfg status

(11) Start ha-lizard - check configuration. All original settings should be restored
service ha-lizard start -w
ha-cfg get


(12) Expose storage on the Master
iscsi-cfg become-secondary (on slave)
iscsi-cfg become-primary (on master)

(13) MIgrate ALL VMs to the MASTER and then repeat steps 4-11 on the slave


(14) Re-enable HA
ha-cfg status

(15) Exit manual mode to restore dynamic storage control
iscsi-cfg manual-mode-disable (on both hosts)

Great SC.

Really thanks for all the details steps.

I will be testing this at the end of month and if you like and keep open the post bring back with feedback.

Thanks!

Hi.
Have you done this? Can you give us a feedback?

Br, Alko.

Hi pudl.

Yes, i test it the steps and at the ends works well. I could upgraded correct both servers.

Hi, pardon the silly question,

But for step 7 "Copy config files from peer to upgraded host" - is the 'peer' referring to the second host, yet to be upgraded?