milenkovicm commented on code in PR #2212:
URL:
https://github.com/apache/datafusion-ballista/pull/2212#discussion_r3706448347
##########
ballista/scheduler/src/config.rs:
##########
@@ -173,6 +173,14 @@ pub struct Config {
help = "Interval, in seconds, to check expired or dead executors."
)]
pub expire_dead_executor_interval_seconds: u64,
+ /// Grace period in seconds to wait for an executor to (re)appear after the
+ /// cluster has lost its last executor before failing the running jobs.
+ #[arg(
+ long,
+ default_value_t = 30,
+ help = "Grace period, in seconds, to wait for an executor to
(re)register after the last executor is lost before failing running jobs.
Prevents jobs from hanging forever when every executor dies, while still
tolerating a transient total loss (e.g. a rolling restart). Set to 0 to fail as
soon as the loss is observed."
+ )]
+ pub no_executors_grace_period_seconds: u64,
Review Comment:
would it make sense to have option to disable this behavior? can we use 0 to
disable it ?
--
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]
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
