Comment 98 for bug 332270
Revision history for this message
| Scott James Remnant (Canonical) (canonical-scott) wrote Re: [Bug 332270] Re: udev repeatedly generates "change" events for the same block device(s) | #98 |
On Mon, 2009-02-23 at 11:32 +0000, MichaelEvans wrote:
> Only, unfortunately, two or more of those are the same real device. A
> race condition (which would probably be aggravated by SMP systems) can
> then occur where a potentially infinite loop of self-chaining change
> events starts. I would presume that at some point it either overflows
> or bogs down to the point where so many instances are spawned that it
> causes the lag and then lets you past.
>
inotify doesn't work this way.
Actually, it looks like something far simpler is happening:
udev runs "lvm vgscan" when it detects an LVM Physical Volume. It looks
like lvm is opening every single block device on the system FOR WRITING;
that triggers the inotify IN_CLOSE_WRITE event, and thus a udev change
event - which means udev runs "lvm vgscan" (repeat until up to
temperature).
It would be helpful if people could reply if THEY DO NOT have *any* LVM
Physical Volumes on their system, and are still experiencing this bug.
Scott
--
Scott James Remnant
<email address hidden>