I have few questions 1. If I have not enabled chec...
# troubleshooting
p
I have few questions 1. If I have not enabled checkpointing in my config, can I still manually trigger a savepoint to upgrade my job? 2. Also, does the restart strategy of
fixed delay
work without enabling checkpoints? Basically, can I define any restart strategy apart from
none
without enabling checkpoint?
m
I'm pretty sure I've seen this question being posted before here on Slack, please don't report
1. If I have not enabled checkpointing in my config, can I still manually trigger a savepoint to upgrade my job?
Yes, see https://nightlies.apache.org/flink/flink-docs-master/docs/ops/state/savepoints/#triggering-savepoints
Basically, can I define any restart strategy apart from
none
without enabling checkpoint?
I believe so, but you will loose your state with every restart
gratitude thank you 2
p
That question was posted by me, but I deleted that, thought I will first introduce myself in the introduction channel and then post the question here.
I believe so, but you will loose your state with every restart
Without a checkpoint, I will lose state with every restart. So thinking of changing the restart strategy so that I can save on container allocation time from Yarn. For context: We are running on Yarn, in per-job submission mode, essentially our job does not share JM and TM.