Андрей Неустроев
Андрей Неустроев
https://github.com/motet-a/jinjalint/pull/33
If everything is ok, I will add changelog and documentations.
> Is the Nomad GC preventing new allocations from being placed until it completes? yes  Why Nomad remove old allocations so long?
sometimes task hang in queue for hours
It's don't depend of client. When it happened I see next log messages. ``` nomad[3402844]:[INFO] client.gc: garbage collecting allocation: alloc_id=1c459fd2-c0e8-d225-5781-96a33c10672d reason="new allocations and over max (200)" nomad[3402844]:[INFO] client.alloc_runner.task_runner: Task event:...
Until all allocs behind `gc_max_allocs` all works as expected. Maybe GC is too slow? or don't clean allocs as expected. server GC settings ``` node_gc_threshold = "24h" job_gc_interval = "5m"...
``` 2024-02-19T03:16:56.261-0500 [TRACE] client.alloc_runner.task_runner: Kill requested: alloc_id=a21ae6da-a252-2f55-a52f-62669c90ca4e task= 2024-02-19T03:16:56.261-0500 [TRACE] client.alloc_runner.task_runner: Kill event: alloc_id=a21ae6da-a252-2f55-a52f-62669c90ca4e task= event_type=Killing event_reason="" 2024-02-19T03:16:56.261-0500 [INFO] client.alloc_runner.task_runner: Task event: alloc_id=a21ae6da-a252-2f55-a52f-62669c90ca4e task= type=Killing msg="Sent interrupt. Waiting 10s before...
Lower `shutdown_delay`, is a solution.