[Ezbench-dev] [PATCH 24/25] runner.py: Signal reboot required on runner error

[Martin Peres]

On 24/02/17 13:19, Petri Latvala wrote:
> Signed-off-by: Petri Latvala <petri.latvala at intel.com>
> ---
>
> This is a hack and a better solution should exist btw. I needed this
> at a time before utils/owatch was able to use hardware watchdogs to do
> a hard reboot, killing the child process group instead. Killing the
> runner was handled in magical ways by the ezbench daemon. I haven't
> been able to even test this path lately.
>
>
> python-modules/ezbench/runner.py | 2 +-
>   1 file changed, 1 insertion(+), 1 deletion(-)
>
> diff --git a/python-modules/ezbench/runner.py b/python-modules/ezbench/runner.py
> index c83bbd2..5d501f2 100644
> --- a/python-modules/ezbench/runner.py
> +++ b/python-modules/ezbench/runner.py
> @@ -395,7 +395,7 @@ class Runner:
>                   cmd_output.append(line.rstrip())
>   
>           raise RunnerError(dict({"msg":"Incomplete command:\n{}".format(cmd_output),
> -                                "err_code":RunnerErrorCode.UNKNOWN,
> +                                "err_code":RunnerErrorCode.REBOOT_NEEDED,
>                                   "err_str":"Stream ended before we got '<--'"}))
>   
>       def __parse_cmd_output__(self, cmd_output):

Yeah, we need a smarter tear-down process, but this patch is really just 
a workaround for it.

I really need to start figuring out what to do there!


Until then, this is a NAK since it is not needed with owatch.