Hi Team, I’m using Prefect server 0.13.10. Sometim...
# prefect-server
a
Hi Team, I’m using Prefect server 0.13.10. Sometimes a scheduled task could exit unexpectedly, and seems Lazarus will re-schedule it after a while. In my case, I want to just let it go and do not want to have it re-scheduled. It’s said that I can be configured per flow in Prefect UI. But what I want is to set this when I create the flow. Is that possible? Thanks.
k
Hey @Alfie, see if you have this in your config.toml and then set it to 0
a
Thanks @Kevin Kho, so the env for this config should be
PREFECT___SERVICES____LAZARUS___RESURRECTION_ATTEMPT_LIMIT
, right?
k
I think that’s right, but not 100% sure it’s available on your version (but I think this should be)
a
Let me try it now, thanks!
k
You might have to restart server Actually that looks like a global setting, let me look for the flow setting
Ignore the above, that’s what I would try
a
Global setting is fine too
In Kubernetes, it’s the env for container prefect-scheduler, right? Or it’s for agent?
k
Both should work I think
a
Thanks. I just set it on the agent, it looks good so far, will keep an eye on it.
Seems it does not work for agent, will try to set it on prefect-scheduler
[2021-09-03T16:05:02.478331+00:00] [info] [7] [140287568955136] [agent] msg="Deploying flow run 7b857363-db38-4a93-bac0-dcdce8139430"
the first time failed, and it’s retried after 10 min
[2021-09-03T16:15:10.798263+00:00] [info] [7] [140287568955136] [agent] msg="Deploying flow run 7b857363-db38-4a93-bac0-dcdce8139430"
k
Ok if that still fails, I think you need to hit the
disable_flow_lazarus_process
graphql API endpoint
a
disable_flow_lazarus_process
works, thanks @Kevin Kho
👍 1