Fix config cannot be updated #5171
Conversation
etam
force-pushed
the
fix-config-cannot-be-updated
branch
from
99becef to
d583b8a
Compare
There was a problem hiding this comment.
I have worries about potential race conditions occurring when the config is updated at the very moment when environment gets enabled/disabled. Environment state could change between status check in update_config and calling super().update_config or assigning _next_config.
|
And you're gonna have a hard time silencing mypy when dealing with class decorators... |
etam
force-pushed
the
fix-config-cannot-be-updated
branch
2 times, most recently
from
f440c41 to
53ff023
Compare
How about now?
It wasn't that bad :) |
scripts/node_integration_tests/playbooks/golem/reconfigure_provider_while_working/playbook.py
Outdated
Show resolved
Hide resolved
| wait_for_computing_task, | ||
| partial(ui_stop, node_id=NodeId.provider), | ||
| partial(change_config, node_id=NodeId.provider), | ||
| partial(check_if_test_failed, node_id=NodeId.provider), |
There was a problem hiding this comment.
shouldn't we somehow detect that the configuration has been completed? (instead of waiting the 10 seconds)?
There was a problem hiding this comment.
I don't know if we should, but in the current "fire and forget" implementation model, it's not easy to achieve. I tried.
Codecov Report
@@ Coverage Diff @@
## b0.23 #5171 +/- ##
==========================================
+ Coverage 89.93% 89.95% +0.01%
==========================================
Files 238 238
Lines 22460 22482 +22
==========================================
+ Hits 20200 20224 +24
+ Misses 2260 2258 -2 |
etam
force-pushed
the
fix-config-cannot-be-updated
branch
from
742fe9c to
cb62998
Compare
Fixes #5151