Using the pod template to configure the local SSD(via host-path or local
PV) is the correct way.
After that, either "java.io.tmpdir" or "process.taskmanager.working-dir" in
CR should take effect.
Maybe you need to share the complete pod yaml and logs of failed
TaskManager.
nit: if the TaskManager pod crashed and was deleted too fast, you could
kill the JobManager first, then you will have enough time to get the logs
and yamls.
Best,
Yang
ChangZhuo Chen (陳昌倬) <czc...@czchen.org> 于2022年4月25日周一 10:19写道：
> Hi,
>
> We try to migrate our application from `Flink on standalone Kubernetes`
> to `Application mode on Flink operator`. However, we cannot configure to
> use local SSD for RocksDB state successful. Any through?
>
>
> Detail:
>
> In original `Flink on standalone Kubernetes`:
> - set `io.tmp.dirs` to local SSD and Flink uses local SSD for its data.
>
> In new `Application mode on Flink operator`:
> - set `io.tmp.dirs` to local SSD causes taskmanager crashloop. We are
> still trying to get the exact error message since it disappers very
> fast.
> - set `workingDir` in pod template does not work. Flink still uses /tmp
> to store its data.
> - set `process.taskmanager.working-dir` does not work. Flink still uses
> /tmp to store its data.
>
>
> --
> ChangZhuo Chen (陳昌倬) czchen@{czchen,debian}.org
> http://czchen.info/
> Key fingerprint = BA04 346D C2E1 FE63 C790 8793 CC65 B0CD EC27 5D5B
>
