F15 blocker? lvm-monitor of snapshot hangs at reboot/shutdown

[James Laska]

On Fri, 2011-04-15 at 15:47 -0400, Clyde E. Kunkel wrote:
 <snip>
> I am wondering if there needs to be new or additional blocker criteria 
> since systemd is now default with F15.  My reasoning is:  with F14 and 
> before, monitoring LVM snapshots was not a problem and shutdowns/reboots 
> simply stopped the lvm2-monitor.  Now since there is a conflict 
> introduced with systemd vs lvm2-monitor (bz 681582), the up-to-5 minute 
> delay in shutdown/reboot could cause users to believe there is a problem 
> that needs fixing and they may spend an inordinate amount of time trying 
> to find the problem.  I looked at the release criteria and didn't see 
> anything that would fit the new systemd scenario.
> 
> I know you indicated in the bz that it qualifies as an Nth for final, 
> but it seems systemd vs the old services start/stop methods calls for 
> some specific criteria.
> 
> What do you think?  Should I just go away and color in my coloring book 
> or is there something here?  Reading the bz, looks like the fix is 
> pretty much straighforward based on Lennart's comments.

Heh, nah ... please share your coloring book! :)  

The good news, you tested lvm-monitor early (e.g. before Final).  As a
result, it seems easily fixed and can be included in F-15 without need
for the blocker/nth process.  Yay, testing early == goodness :)

As far as whether we should have criteria that cover LVM snapshots, that
feels like it falls under local site configuration [1], and we'd need to
consider several things like # of impacted users etc...  

With regards to the timeliness of the reboot and shutdown operations,
perhaps there is something there we can clarify.  For example, if
reboot/shutdown was taking 5 minutes for a default installation and
impacting *all* users, my spidey sense makes me believe that would be a
valid blocker.  It could be argued that a 5 minute reboot/shutdown does
not honor the current Beta criteria ... 

        "The desktop's offered mechanisms (if any) for shutting down,
        logging out and rebooting must work"

But that's a bit open to interpretation.  For example, it could be said
(I think I even did) that reboot/shutdown does work, it just takes
forever.  Anyone else think it would be worthwhile to adjust the Beta
criteria to something like ...
        
        "The desktop's offered mechanisms (if any) for shutting down,
        logging out and rebooting must work and perform the intended
        function in a reasonable time."

Thanks,
James

[1]
https://fedoraproject.org/wiki/Blocker_Bug_FAQ#What_about_hardware_and_local_configuration_dependent_issues.3F
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 198 bytes
Desc: This is a digitally signed message part
Url : http://lists.fedoraproject.org/pipermail/test/attachments/20110418/9820aa79/attachment.bin