Sometimes, our users are running into problems during boot where devices are not properly set up and given that event-based activation takes place, it's even more important to collect coherent set of information from single boot sequence.
Asking for more and more information gradually is not quite effective as device names can change after reboot and then we need to gather full set of debug logs again. This can be very time consuming while waiting for user's output. It's very easy to forget one piece of log that may be important and then we can end up asking for the same set of debug info again + something more.
This report is just to make this debug gathering to be more automatic where, in ideal case, a user need to run as least commands as possible (in very ideal case, just 1 command/script to set debug mode for next boot sequence and then chaning it back automatically) - so instead of asking users to run concrete commands step by step to set the state of the system for it to be prepared for gathering various boot-related debug logs (also often requiring a need to modify existing scripts etc.), we should make it all prepared in a way (from LVM point of view) that using a kernel command line option would switch all the parts we need during boot into debug mode and then gathering this information automatically into a single dump that the user can send us for inspection.
See also bug #1287106 comment #20 with an example of steps needed to get complete log about events and then associated system state to make progress with inspection of the problem. Investigate if there's anything more needed.