lostluck commented on code in PR #28545:
URL: https://github.com/apache/beam/pull/28545#discussion_r1331813802
##########
sdks/go/pkg/beam/runners/prism/internal/execute.go:
##########
@@ -95,7 +92,7 @@ func makeWorker(env string, j *jobservices.Job) (*worker.W,
error) {
// Check for connection succeeding after we've created the environment
successfully.
timeout := 1 * time.Minute
time.AfterFunc(timeout, func() {
- if wk.Connected() {
+ if wk.Connected() || wk.Stopped() {
Review Comment:
No. The stopped bit is only set when Stop is called, which only happens when
a context has cancelled, which only happens on post Job clean up.
Other errors would be reported earlier (ideally).
This approach largely prevents this ("worker didn't connect") function from
failing successful jobs after the job completed, if the job never needed the
environment at all. I intend to change worker environments to start on demand,
and move worker startup to after preprocessing. As it stands, a minute+ job
using one of these will just die. Won't happen for real xlang transforms though.
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]
