Nope, the servers that the Beta software runs on are not mission critical at all - that is the point of beta testing [[:)]] I tend to run some production grade traffic through it to give it a workout but I also have GA release Astaro servers that are taking the load. As a fellow regular poster I'm sure that you've seen me post before in support of the terrific work that the Astaro team does, and, also my comments to others about not putting beta release software into a production state environment unless you're prepared to accept the potential risk!
However, as much fun as testing beta software is, there do seem to be some defects that have got into this release - which affected a LOT of beta testers, which with just a little more testing could have been spotted. Perhaps I had been drinking a little too much coffee today [[:)]] so my apologies if my comments were construed as a little sharp.
As for not updating all nodes within a cluster at the same time - Astaro design is that for an active/active cluster (which basically turn multiple servers into one big server) all nodes must use the same Astaro release and therefore when applying an update to a cluster, it is first automatically applied to the slave nodes, which are then rebooted, synced and return back online before it is automatically applied on the master node.
Nope, the servers that the Beta software runs on are not mission critical at all - that is the point of beta testing [[:)]] I tend to run some production grade traffic through it to give it a workout but I also have GA release Astaro servers that are taking the load. As a fellow regular poster I'm sure that you've seen me post before in support of the terrific work that the Astaro team does, and, also my comments to others about not putting beta release software into a production state environment unless you're prepared to accept the potential risk!
However, as much fun as testing beta software is, there do seem to be some defects that have got into this release - which affected a LOT of beta testers, which with just a little more testing could have been spotted. Perhaps I had been drinking a little too much coffee today [[:)]] so my apologies if my comments were construed as a little sharp.
As for not updating all nodes within a cluster at the same time - Astaro design is that for an active/active cluster (which basically turn multiple servers into one big server) all nodes must use the same Astaro release and therefore when applying an update to a cluster, it is first automatically applied to the slave nodes, which are then rebooted, synced and return back online before it is automatically applied on the master node.
