I 'THINK' this is a udev problem?[Solved]

[joe3]

I just did a somewhat belated "equo update && equo upgrade --ask", I've been having trouble getting time to play with my desktop PC lately, so this is the first time I've updated my desktop's Sabayon since just before Thanksgiving.
Though I have been getting around to updating my laptop a bit more often...

Anyway when I rebooted the desktop, I noticed a longer delay than usual when tty1 displayed this:

Code: Select all

* Starting udev ...                                                      [ ok ]
* Populating /dev with existing devices through uevents ...              [ ok ]
* Waiting for uevents to be processed ...


When it eventually gives up I get:

Code: Select all

udevadm settle - timeout of 60 seconds reached, the event queue contains:
 /sys/module/nvidia (6369)
 /sys/module/nvidia (6372)                                               [ !! ]


The next thing I noticed was that when I logged in to my normal user account, my personal data partitions failed to mount:

Code: Select all

localhost login: jtwdyp
Password:
Last login: Thu Jan  3 19:02:30 EST 2013 on tty1
You have old mail in folder /var/mail/jtwdyp.
mount: LABEL=jSTUFF_desk-11: No such file or directory
mount: LABEL=Jimages_desk-9: No such file or directory
mount: LABEL=J10_desk-10: No such file or directory


The said named partitions do still exist:

Code: Select all

JtWdyP -> /home/jtwdyp
> ls /dev/disk/by-label/[jJ]*
/dev/disk/by-label/J10_desk-10     /dev/disk/by-label/jSTUFF_desk-11
/dev/disk/by-label/Jimages_desk-9
JtWdyP -> /home/jtwdyp
>


My .bash_profile used to mount them like so:

Code: Select all

JtWdyP -> /home/jtwdyp
> grep mount ~/.bash_profile
mount LABEL=jSTUFF_desk-11
mount LABEL=Jimages_desk-9
mount LABEL=J10_desk-10
JtWdyP -> /home/jtwdyp
>


It relied on these fstab lines:

Code: Select all

JtWdyP -> /home/jtwdyp
> grep owner /etc/fstab
LABEL=jSTUFF_desk-11 /home/jtwdyp/STUFF/desk-11 ext3    defaults,owner,noauto      0       0
LABEL=Jimages_desk-9 /home/jtwdyp/images/desk-9 ext3    defaults,owner,noauto      0       0
LABEL=J10_desk-10 /home/jtwdyp/desk-10 ext3    defaults,owner,noauto        0    0
JtWdyP -> /home/jtwdyp
>


Which worked because of this rule:

Code: Select all

JtWdyP -> /home/jtwdyp
> cat /etc/udev/rules.d/99-z-local-disk.rules
# /etc/udev/rules.d/99-local-disk.rules
SUBSYSTEM=="block", ENV{ID_FS_LABEL}=="jSTUFF_desk-11", OWNER="jtwdyp", GROUP="jtwdyp"
SUBSYSTEM=="block", ENV{ID_FS_LABEL}=="Jimages_desk-9", OWNER="jtwdyp", GROUP="jtwdyp"
SUBSYSTEM=="block", ENV{ID_FS_LABEL}=="J10_desk-10", OWNER="jtwdyp", GROUP="jtwdyp"
JtWdyP -> /home/jtwdyp
>


Anybody have any idea what went wrong? Or better yet, how to fix it?

--
Joe3

[Fitzcarraldo]

Looks like udev is having trouble loading a module. Have a read through the thread Long pause at 'Waiting for uevents to be processed' [Solved] for the method of investigating and find the module that is causing the problem.

[joe3]

Looks like udev is having trouble loading a module. Have a read through the thread Long pause at 'Waiting for uevents to be processed' [Solved] for the method of investigating and find the module that is causing the problem.

OK so there were two problems...

the nvidia driver wasn't loading. I couldn;t figure out why, so since I was already limited to console without it, I used equo to reinstall the same nvidia driver that equi search said was installed. During the reinstall, there was some warning text about inst6alling a driver known not to work with an installed video card. (there is only the built in card, and I wonders what nvidia driver I was using before the last equo upgrade... Anyway since that didn't solve it I removed nvidia-drivers-310.19 and installed nvidia-drivers-304.64... Which downgraded something called userspace, and upgraded the kernel from x86_64-3.3.0 to x86_64-3.7.0

I no longer get that udev timeout on boot...

But my .bash_profile still wasn't mounting my data partitions...

At this point I looked a little closer at /dev/disk/by-label, and while ls -l reported the ownership of my partition device names as "root root", when I was about to try using ^Co to try chowning them from a root midnight commander session, mc's chown pop-up reported then as already belonging to "jtwdyp jtwdyp"
So while it bugs me a little that ls -l doesn't show the ownership asigned by my udev rule, it does seem that the rule is still being processed.

So I tried editing my .bash_profile to use:
mount {fstabListedMountPoint} IE "mount $HOME/images/desk-9"
instead of the:
mount LABEL={fstabListedSeviceLabel} IE "mount LABEL=Jimages_desk-9"
{which used to work...}

Using the mountpoint instead of the device label in the mount command seems to work.

But if udev is reassigning the partition ownership according to my rule, shouldn't
"ls -l" show it???