| Summary: | udev not creating /dev/sdxX devices for some partitions | ||
|---|---|---|---|
| Product: | Mageia | Reporter: | Frank Griffin <ftg> |
| Component: | RPM Packages | Assignee: | Mageia Bug Squad <bugsquad> |
| Status: | RESOLVED FIXED | QA Contact: | |
| Severity: | normal | ||
| Priority: | Normal | CC: | mageia |
| Version: | Cauldron | ||
| Target Milestone: | --- | ||
| Hardware: | x86_64 | ||
| OS: | Linux | ||
| Whiteboard: | |||
| Source RPM: | udev | CVE: | |
| Status comment: | |||
|
Description
Frank Griffin
2013-04-29 21:13:48 CEST
claire robinson
2013-04-30 10:27:07 CEST
CC:
(none) =>
mageia I *really* don't think it's related to fstab! Can you confirm this however? Summary:
udev not creating /dev/sdxX devices for partitions not in fstab =>
udev not creating /dev/sdxX devices for some partitions I saw this in two scenarios. I keep three root patitions on the HD and rotate them for fresh installs and fallback. They have been there, formatted, forever. The ones other than the booted root are not mentioned in the fstab of the booted one, and that is exactly the subset for which devices weren't created. In the second scenario, I created a new install root partition, did an install to that, and rebooted the normal system. The new partition was also not in fstab, and did not have a device created. I then added all of the devices to fstab, rebooted, and devices were created for all of them. I can't tie it up any tighter than that. Another interesting, although unrelated, issue is that if any of the partitions in fstab are not available at boot, systemd drops you into a recovery shell. I found this by adding some partitions on a flash drive and then booting without the drive plugged in. Very interesting indeed - I may need to eat my words in the last comment! :) I will have to try and recreate this to debug it further I think. As for the recovery shell thing, yes, this is expected. If you mention a filesystem in fstab where the device does not exist, systemd will consider this a critical part of the boot process and drop you to shell to fix it. If you don't care about such mountpoints (i.e. you as the administrator feel they are not important or critical to the boot), you can mark them with the nofail option and systemd will continue the boot as normal. I think this behaviour makes sense generally (e.g. you likely wouldn't want the boot to continue with your /var partition not mounted for example, but your /data partition might not matter too much in your setup so you can configure them accordingly to your needs). I tried to reproduce this in a VM. I created partitions / /foo /usr /bar swap in the installer and then removed the mountpoints for /foo and /bar. The install when fine. After rebooting, I could see all my partitions fine. I then dd'ed them to 0 and rebooted, but again they still showed up. (they are of course unlisted in fstab). So I'm really struggling to see how to reproduce this issue. I wonder if there is some kind of issue with udev itself? Perhaps the recent update to fix the FPE (floating point exception) stuff properly has actually fixed this problem? I'll retest on the affected machine and let you know. You're right. In current cauldron (as of about 10 mins ago), the missing devices are now being created just fine. Status:
NEW =>
RESOLVED Nice thanks for confirming :) |