問題文
A security fix has been applied inside the software image used by 30 running compute nodes. Jobs are running on all of them and the operator wants the updated files on the nodes with as little disruption as possible. Which approach fits the requirement?
選択肢
- Copy the changed files from one running node back into the image, which distributes them to the remaining nodes as a side effect of the next configuration sweep.
- Reboot every node, because a reboot is the only path by which any change made inside a software image can ever reach the nodes that use it, whatever the nature of the change.
- Push the image changes to the running nodes with the image update operation, which synchronizes changed files onto a node that stays up.
- Edit the files directly on each node over a remote shell, then commit the result so that the image picks up the same edit and later reinstalls carry the fix forward.