We have a config cluster with nodes with a higher amt of latency. I dont want to wait for the upgrade to get pushed from cluster members. Can I start the upgrade on all members at the same time?
We have plenty of nodes are are taking the affected cluster out of service so no users will be on the affected boxes.
I do not know if I understood your question correctly, but I can share what I know.
If you have a cluster, you can not start upgrading all the nodes at same time, because when you start an upgrade process in a cluster, it will upgrade a node each time. If you do that (all at same time) you will stop serving your customers somewhere in time while the nodes reboot and start the upgrade process. The only way I know to do that, is destroy the cluster, upgrade all the nodes at the same time, them, recreate the cluster.
One action you can speed up, is the availability of the package in all nodes prior to the update. To do that, you can use the staging area.
Thank you for the response!
So we have multiple multiple seperate clusters and use DNS to control what boxes are able to serve clients. Assuming we dont care about all cluster nodes going down at the same time, would we be able to just reboot all cluster boxes at the same time without breaking anything?
Thank you again for the help!