Using UUID in /etc/fstab causes long delay on boot

Bug #66170 reported by John S
2
Affects Status Importance Assigned to Milestone
linux-source-2.6.17 (Ubuntu)
Invalid
Undecided
Unassigned

Bug Description

When any lines in /etc/fstab use the UUID= method to select volumes, there is a several-minute delay at boot, during which the splash screen disappears and the screen is blank except a blinking cursor. During this time, all drives display regular flickers of activity several times a second.

All UUIDs are accurate and up to date. If no UUID specifications are in fstab (/dev/hdx instead) then booting occurs normally. There are NTFS filesystems in fstab, but it doesn't have to be these that use UUID specification in order to see the problem.

Revision history for this message
John S (jcspray) wrote :

Bizarre -- now the delay is happening again without any UUID specifications in fstab. Red herring I guess.

Revision history for this message
Fabio Massimo Di Nitto (fabbione) wrote :

Can you please check if you have installed mdadm or lvm-common? if you don't need them at boot, can you please remove them and see if the timeout still occours?

What kind of devices do you have connected to your system? usb? are they "slow" to start up?

What about changing root=UUID to root=/dev/.. in grub?

Fabio

Revision history for this message
John S (jcspray) wrote :

Curiouser and curiouser...

I tried remove mdadm, lvm-common and lvm2. They rebooted and hooray, the wait is gone. They, in attempting to isolate where the problem was, I addeded them back one by one. Eventually, I had reinstalled all three packages, and the wait had still not come back. Something in the install/uninstall process I suppose?

To answer the rest of your questions -- the devices are /dev/hda (internal IDE) and /dev/sda (external USB), although the USB drive is not normally turned on at boot.

The root device is specified to grub by UUID, I haven't tried changing this since the system appears to have unbroken itself anyway.

Revision history for this message
Kjell Braden (afflux) wrote :

I'm closing this bug as it sounds like you don't experience it anymore. In case you still do, feel free to reopen this bug.

Changed in linux-source-2.6.17:
status: New → Invalid
To post a comment you must log in.
This report contains Public information  
Everyone can see this information.

Other bug subscribers

Remote bug watches

Bug watches keep track of this bug in other bug trackers.