Closed carpawell closed 7 months ago
If something fails
Then it should be zero exit code. If it exits with non-zero code, nothing changes with this PR.
not masking the behavior with restart (which if OK for production).
Do not see any differences. In both you should go with caution, you have logs, you have docker stats (uptime), healthchecks.
I do not insist, but on-failure
is senseless to me (moreover, not every service has it).
Maybe it should be "never" then?
Maybe it should be "never" then?
That is better to me than on-failure
(if every service has it). At least that means something.
@roman-khimov, let's do something here. never
and merge or just closing?
I'd ask @cthulhu-rider first of all. IMO never
is the best option, but I'm not using dev-env that often.
I suffer from such restarts as well. After laptop restart dev-env in broken state and you have to drop and start again
Changed to never
, @all, make your choices.
Well, not exactly that all...
@nspcc-dev, @nspcc-dev/neofs-core, @nspcc-dev/neofs-services
btw is never
a correct setting? https://docs.docker.com/compose/compose-file/compose-file-v2/#restart
seems like it should be no
which is default
If a user runs services in the devenv it is most likely that they want it to work and serve for some purposes trying their best. Unconditional restart is also smth that you usually can see on a real machines.