Yeah I was a bit disappointed with that too. I have a script that runs every morning and updates all my nodes with a delay of 4 hours between each node, starting with the smallest node. I have 6 nodes.
4 nodes got updated before the ingress was gone so the last 2 nodes suddenly got no ingress. I was like: “wth… the image just got pushed. It’s not even 3 days yet so most watchtower instances didn’t even get a chance to update.”
Oof yeah that’s cutting it pretty close. I don’t even think the 3 day docker update period has ended yet… That doesn’t seem right. I think someone forgot 1.27 was skipped to be honest. This seems a little unfair to people who even have watchtower running. I updated manually yesterday, but going by my watchtower logs, my nodes wouldn’t have been updated yet.
Missing half a day of ingress is not a big issue, but nodes that are 3 minor versions below latest are disqualified, so this update put all docker nodes 1 version away from DQ. Storjlabs please follow your own bi-weekly release schedule or at least don’t bump minor version before everyone get a chance to update.
Funny how if you hover over the version number v1.28.2 in the dashboard it says “Running the minimal allowed version: v1.24.0”. I wonder if this is hard-coded and just plainly wrong?
Guess I could try and read the source code to find out if I could be bothered.
We are going to start disqualifying nodes that are 3 minor releases behind. In the above example, 1.10 and older will get DQed. We will also be ensuring that the standard source code release will simply refuse to start if it’s 3 minor releases behind.