TOPIC:

Hi, I read past posts on which there were procedures to follow in order to perform a correct upgrade of XenServer- XCP-NG without creating problems to the halizard - no san and mantain the uptime of the machines.
I'm wondering If these steps are still valid considering that on a most recent post there were the reference to rpm packages or if there would be different steps to perform with the most recent version of the haliozard nosan package.
I did not find any further indication apart the posts and also the instructions does not mention any rpm.

Thanks a lot, kind regards.

Giorgio

The most recent post which discusses the rpm packages is the one you should follow. The process is simpler and the latest release adds new functionality which stores all lvm, drbd, tgt params in the pool DB, making it now possible for the full configuration to survive a major release upgrade of xenserver or xcp-ng.

Bare with me but I'm a bit confused between the installation video and relative instruction for the AIO package of the nosan installer and the rpm you talk about. Do you mean that the instructions are still valid but i do have to simply install rpms that include a yum mirror or whatelse and follow the instructions of the video? Where can I find thos rpms?
Kind regards

Giorgio

The installation script and video are still valid and the best way to deploy a new hyperconverged 2-node cluster. It will automatically install all dependencies and the latest versions of our code. It will install the same version as the rpm we discussed in the earlier post. This script/prodedure should not be used for upgrading an existing cluster.

There are 2 types of upgrades.

1) halizard/iscsi-ha. This is simply an upgrade to our SW. In the past, users would download a tarball of a new release, unpack, and then run an installation script. Although this is still supported, we have moved to rpm packaging and the preferred method is to upgrade via rpm.

2) major upgrade to xcp or XenServer. In this case, our installation gets wiped by the new o/s install. In the past, we have published instructions for performing such an upgrade with zero downtime of the VMs. That process had many steps and is still valid. However, our latest release, 2.2 (which coincidentally is our first rpm based release) has a number of new features which makes upgrading dom0 much easier. Most important of these features is that we now store ALL configuration data (drbd,lvm,tgt) in the pool DB, so, after upgrading dom0 all of your settings are preserved. Namely, all that is required to get the cluster going again is installation of the dependencies and restoration of the config. This is the upgrade process you referenced a few days back from another post.

Regarding where to get the latest versions. All versions are available here:
halizard.org/release

Ok then I'll proceed with the script as a very first step.