All my nodes are running storagenode/beta since jan 2020. This week i’ve setted up new storj miner for my friend, and noticet that storj provides startup command with storajenode/latest. I’ve tried to restart mine node, previusly running on beta, using latest container, but it don’t start. So i moved back to /beta. Is there any penality for using /beta, or that ok and i’m shouldn’t worry about it?
You can read more about it here The storagenode docker run command has been updated
The two images currently seem to be the same so there shouldn’t be any difference if you run one or the other. Anyway you should switch to latest, to be safe if this will change in future.
Did you try to stop/remove and restart the new image?
I know back when I was trying to switch this as well, I had to stop and rm the container, but then also run “sudo docker system prune -a” to totally remove the old image, then when restarting with the updated command, it will download the “latest” image and start right up.
I believe it will use the “latest” tag going forward if you just rm and recreate. Not super necessary to prune. Watchtower uses the tag in the run command when it pulls new images.
always do the docker rm storagenode when you manually update, else the update will not always work without a restart or something similar to purge the loaded image from docker.
else beta was the old “tag” which everybody used, so it will most likely keep being updated for a long time… but one should really switch to latest, even if there is no good reason to do so, atleast that i’m aware off…
there shouldn’t be any different in the versions, i just waited until there was a new version and then forgot to go to latest, but then next time i remembered… but really no big point in that…
the image is downloaded online and used as a virtual harddisk for the storagenode, so really nothing of it really matters… only thing that really matters is that you never go backwards in versions…
there is no telling what kind of incompatibility issues that could cause if done now or in the future.
but yeah going from beta to latest is like a 1 minute thing per node… and really straight forward
just remember doing the rm storagenode and having your run command ready
you might remember it, else you can find the syntax here… and ofc make sure everything is right
Perhaps you forgot to pull the image
docker pull storjlabs/storagenode:latest
Without it the docker will use a downloaded version of the latest image on you PC, which more like outdated. It doesn’t check updates for the
:latest images by default in the
docker run command, if this image is already downloaded. This is downside of any static tag, either
latest. So you always should explicitly pull the image before the run.
Yea, i know that. So if i check…
Yep, it is updated
Then it should work. If it’s not - please, post an error
Same situation here. Despite (trying) getting latest version v1.13.1 with the
docker pull storjlabs/storagenode:latestit remains stuck on v1.12.3. Not sure the reason, nor how to fix it?
Welcome to the forum!
Do you have a error too? Which one?
If only not the rolled version for docker, then it’s expected:
Thanks fir the swift reply @Alexey. Looking at change log on GitHub and seeing the new version 1.13.1 be published, I went for a manual update using the docker instruction to proceed. Is it possible that docker does not download the version marked as « latest » on GitHub, upon manual action to specifically pull the latest? Btw, believe i was not in a beta before, I have been running exclusively release versions, hence not really matching this thread’s situation
i just checked, the version on the docker image is 12 days old, so that would be v1.12.3.
the v1.13.1 version is released for the windows clients, and takes like a 7 or 8 days from the windows release until the docker version is release
so yeah a storagenode still running v1.12.3 or downloading it…
well it’s not a bug… it’s a feature
why is it a feature, well because if one wants to keep a network working 24/7/365
then running one version means 1 error can drop most of the network…or it’s more likely to … the time gap in version releases is a failsafe to help ensure network stability, even if all else fails.
maybe it also helps debugging code…
here you can check the images here
thanks @SGC, glad to know and fully understood. let’s be patient
i should also recommend that you should use watchtower… but i don’t so…
not a fan of automatic updates for critical components
but i’m sure it’s very good and makes life much easier
I have watchtower activated, as per following the instructions on how to setup a node, official mode. Looking forward to release on docker!