Good thing you’re not brave enough as that certainly won’t work. It’s not an actual release and the node likely wouldn’t start. Always use the image mentioned in the docs.
well some people have gotten the v1.1.1
has to had come from somewhere… just saying…
The windows nodes are updating at the moment. We have plans to move the docker nodes to the same update process but for the moment they are disconnected and because there are still nodes out there that don’t follow our recommendation we push the docker image last. Just for safety.
My docker (2 Nodes) updated 5 h ago.!
Please make sure you use the recommended watchtower version. If both nodes update at the same time (within a few minutes) that indicates you are not following our recommendation.
2 Nodes on different locations and watchtower by storj (new version)
I tried. On a node that runs 1.0.1. For those that would like to try - It didn’t break the data but also didn’t start so don’t do it:). Pulling a newer image on :beta
did’t give me the newer version so we need to wait patiently.
Some more feedback after receiving the update. I guess I’m crashing the party too.
- On the all satellites screen it seems the held back amount is not shown, but it is incorrectly subtracted between Gross Total and TOTAL. It assumes 75%, just like what is displayed on the separate satellite pages. I would suggest definitely displaying this line with percentages omitted as they can differ per satellite. Otherwise it’s really hard to find out why the TOTAL is suddenly 75% lower.
- The Estimated Gross Payout seems to be exactly double the TOTAL in my case. Perhaps that’s just the timing, that wasn’t exactly the case when @littleskunk posted his feedback. I think this was actually an attempt to calculate expected payout by the end of the month, but the calculation is simply off. We’re nowhere near halfway through the month, so it should be closer to 4x TOTAL.
- << It’s still incorrect like all held amount related data, but that’s not what I want to point out. It should be “1st month” or “month 1” if you don’t feel like programming around the 1st, 2nd, 3rd etc. postfixes. Additionally since this counter applies to this specific satellite and not the whole network, it should probably be “on this satellite” instead of “on the network”.
Other things were covered by @littleskunk already.
Is that how it is supposed to work now?
My watchtower updates itself and all my nodes (and a few other containers ) and it always updates all 3 nodes at the same time.
The statement that it’s my first month is not true. I’m now in my fourth month. Started on 3 January.
Edit: Just noticed that BrightSilence already showed that mistake
That means you are running a different watchtower version. The latest version should update over a periode of 3 days. They shouldn’t update all at the same time.
I would like to point out one more thing. In web ui last contact shows “0m ago” on enter and after that starts counting time. When you refresh it resets counter and shows “0m ago” again.
CLI dashboard with newest updates shows ONLINE next to Last Contact (this one thing is two different things in web ui - state + last contact). It would be good to have cohesive vision on that and make design decision “Does user needs to see last contact or no”. Maybe for web ui user and cli dashboard it is only important to show current state -online or not. And something with more detailed thing like rest api could have additional info on that. As last contact functionality was cut out of latest cli dashboard so shouldn’t “Last Contact” be renamed to STATUS as this is the name for this kind of functionality in web ui. This can be confusing for new comers.
I use storjlabs/watchtower:latest
It does update over 3 days in a random interval, but once it found a new image, it updates all containers running that image.
Yeah this happens to my nodes, they both update at the sametime on the same machine.
Oh I see. Yes that makes sense.
It’s fixed in the v1.11.1