Closed kke closed 6 months ago
I used k0sctl v0.17.8 and I confirmed that all workers cordoned before upgrading.
Why does you create another BatchedParallelEach? @kke https://github.com/k0sproject/k0sctl/pull/672/files#diff-1eafa28461253d6be29ff9ff2ef59982fe5193b2cdc0f507e68a56683c47ec36L90
Why does you create another BatchedParallelEach?
Yes I don't see any reason why it couldn't be in the existing one looking at it now with fresh eyes.
Fixes #670
There was a mistake in #595 that made worker upgrading cordon all the workers before an upgrade. The intention was to cordon only the current batch, so that workloads don't migrate to the nodes in the same batch.