-
|
We currently have a 3-nodes etcd cluster running on our servers. It is used as a distributed configuration store for multiple (thousands) PostgreSQL databases using Patroni for high availability. These databases use various Patroni versions, from 1.6.5 to 2.1.7. Most of these databases (but not all of them) use etcd API v2. Each one of these PostgreSQL databases have a specific username/password to connect to etcd. We currently use etcd v3.4.16 and plan to upgrade to 3.5.7. We noticed in the etcd documentation (https://etcd.io/docs/v3.5/upgrades/upgrade_3_5/) the following note in the "Upgrade checklists" section:
To make sure I correctly understand, this means that there is no way for us to do a rolling upgrade of the etcd cluster? In this case, is the solution to stop the three nodes, upgrade etcd, start again the three nodes? Thanks for your answers :) |
Beta Was this translation helpful? Give feedback.
Replies: 1 comment
-
|
For the record, we upgraded our etcd clusters in a rolling upgrade manner with no visible impact despite the fact that authentication is enabled. Hence I don't know what this note in the etcd documentation means but it does not seem to be relevant. |
Beta Was this translation helpful? Give feedback.
For the record, we upgraded our etcd clusters in a rolling upgrade manner with no visible impact despite the fact that authentication is enabled. Hence I don't know what this note in the etcd documentation means but it does not seem to be relevant.