Improve restart of pmc montoring #113
Conversation
|
Thanks for your PR, |
pkg/daemon/daemon.go
Outdated
| go pmc.Poll() | ||
| for { // Wait for pmc.parentDS to be updated by Poll | ||
| time.Sleep(10 * time.Millisecond) | ||
| if pmc.parentDS != nil { |
There was a problem hiding this comment.
If RunPMCExpGetParentDS fails, this will never exit this loop?
There was a problem hiding this comment.
Even if the poll fails the expect should pick up a change at some point so it might be delayed but should exit in decent time pratically.
I can add a timeout if you'd like I'll just make sure its so long we are unlikely to hit it.
There was a problem hiding this comment.
Also for context this function is called when cloud-event-proxy start/restarts to emit the clock class. So either we're starting up and will receive/update in not long or will already have it so won't enter this block.
There was a problem hiding this comment.
I've realised I don't actually need the poll here so I just removed it and left a comment on why we can skip.
The polling interval was cause delays in killing processed My moving into a go routine processes it no longer blocks actions such as stopping the process.
nocturnalastro
force-pushed
the
pmc_monitor_process_with_goroutine_expect
branch
from
eb0b789 to
ea5deff
Compare
2 participants
The polling interval was cause delays in killing processed by moving into a go routine processes it no longer blocks actions such as stopping the process.