Closed wwestgarth closed 2 years ago
I was working with a few people and I found an interesting thing. Let's say We have the following use case:
PyCharm
nomad
in terminal 1
, and keep this open foreverterminal 2
, or somewhere... Wwhen the PyCharm
is restarted for any reason (like: plugin upgrade, crash, by users mistake...), the nomad is killed but network processes are orphaned. It means you cannot start a new network because old programs are running and you can see: Cannot bind-address: 0.0.0.0:XXXXX. Port already in use
I think We should add a flag for this command(maybe --kill-processes
) to kill all processes. Names for processes We should take from the configuration.
I just did a workaround for the system tests, but it may be done better here: https://github.com/vegaprotocol/system-tests/pull/595/files#diff-8b119a9bacccbd06b544bc77467bcd92e68bd86f980367cb45d4379496759f46R246-R251
Close with #59
Being able to start and stop a network is amazing, but both tendermint and vega's state persists. This means that when you restart the network the vega core will replay all the blocks. Sometime's, though, you just want to start from fresh again (without having to do the lengthy destroy/bootstrap).
It would be useful if there were a
vegacapsule state unsafe_reset_all
or something that when the network is stopped, will go and callvega tm unsafe-reset-all
andvega unsafe_reset_all
for all the nodes.