We am currently planning a migration of our 6 Windows Server 2008R2 Connection Brokers to Server 2016. In the below KB it mentions that the maximum number of Broker servers in a pod is 7.
Just wondering if anyone as run 8 brokers in a pod and had any issue or did it just run as expected? The reason I ask it would make the migration that little bit easier with change\outage windows for the team. For us it would only be a transitionary state for maybe one day not a permanent configuration.
Open a GSS ticket, but my understanding its a performance issue due to replication.I believe there are some options to push the limits that are done in larger enviornments, but before you think of trying anything I'd make sure its supported and you have that documented in a ticket,.
Limitation is due to JMS traffic between brokers, I'd suggest to upgrade brokers one by one by provisioning new VMs with Windows 2016, then install Horizon on of them to integrate the existing cluster and remove any of the old Horizon CS and repeat until you replaced all older VM
Thanks @EricMonjoin Yeah is was sorta assuming that was the case with the JMS traffic 😔😕
That is our plan to swap out one by one. But was looking for a shortcut due to our implementation complexity I wanted to keep the outage as short as possible. Just wanted to see what the greater Horizon community would say, if we where to burst greater than the 7.