On Tue, Dec 3, 2019 at 3:29 PM pmkellly@frontier.com <pmkellly@frontier.com> wrote:
On 12/3/19 06:27, Kamil Paral wrote:
> On Mon, Dec 2, 2019 at 7:43 PM pmkellly@frontier.com <pmkellly@frontier.com>
> wrote:
>
>> I have the update ready. I think I interpolated the discussion
>> correctly, but probably leaned toward the "keep it simple for now" case.
>> Please let me know if there are further changes needed. Here's the link:
>>
>> https://fedoraproject.org/wiki/User:Tablepc/Draft_testcase_reboot
>
>
> I was hoping you would include `journalctl | grep` commands that would look
> for the strings Chris mentioned, so that it's easy for people or for
> scripts to confirm whether the filesystem was or wasn't properly unmounted.
> It's still helpful to include the examples (they don't necessarily need to
> be Expected Results, but it's fine there), so that people can compare the
> full text if they want or have doubts, but those grep commands would make
> the comparison much simpler.
>

I took a quick tour back through some the the recent e'mail, but didn't
see any thing about grepping the journal.

Here's the email I had in mind, containing the important journal messages:
https://lists.fedoraproject.org/archives/list/test@lists.fedoraproject.org/message/S5LQTXKFPIUKBFC7HL5SNSAV3ZRFP6WV/

We need to create a command or commands that will detect any of the failure states (and ideally another command for any of the success states, for confirmation).
 
From the above, I'm thinking
it might be good to do the journalctl -b > journal.log first and then
grep the file for the unfortunate result phrases. Then if one of the
phrases is found in the file ask the tester to file a bug report with
the journal file attached. Is that what you had in mind?

It doesn't really matter if you grep the journal directly or save the journal and grep the file. But it's one fewer command to grep the journal directly. And if a failure is found, ask the user to save the journal and report a bug.