I had built myself a chroot environment on to run 32 bit apps on my 64-bit ubuntu box...I did a bonehead thing in that I tried deleting it while it was in use.

Now I need advice on how to rebuild /dev. Fortunately, the machine is still running and I've restored my home dir, and many of the nodes...but I want to be *really* sure before rebooting.


I have done the following:

1) Run all of the MAKEDEV scripts

2) run mknod <devname> <type> <major> <minor> for every device I could think of and find in the devices.txt file

3) Listed all of the devices in /proc/devices...

4) Now what?


My biggest worry is /dev/mapper I have no idea what was in there or how to rebuild it.

Tips? Advice? General taunting? Constructive help preferred.

You don't use udev or devfs?

I'm not afraid of config files et. al. but after a moderate amount of homework, I'm not sure what I'm looking for.

I found /etc/udev and see that I have plenty of rules and whatnot in there. I found the following in udev.conf, but not the associated file:

udev_db="/dev/.udevdb"

Is this another item I will have to rebuild?

I also did "udevinfo -s" and it returned nothing.


Can you point me in the right direction?

Run 'ps -ef | grep udev'. If udevd is running, then you SHOULD be able to reboot without a problem. udevd only requires /dev/console and /dev/null to actually exist: it will recreate everything else (or should).

fld094, are you using the 2.4 or the 2.6 kernel? If I recall correctly, 2.4 does not use udev.

Ok...just double checking on a few things because I'm paranoid:

1) udevd is running and I am running a 2.6 kernel...so it should remake the "device-mapper" bits automatically? When I run MAKEDEV update, it complains about nearly everthing.

2) I am running raid 5 with 4 disks...(md0) Is there anything special I should have to check for them? Here is the output from "mount"

/dev/md0 on / type ext3 (rw,errors=remount-ro)
proc on /proc type proc (rw)
devpts on /dev/pts type devpts (rw,gid=5,mode=620)
/dev/mapper/tmp-tmp on /tmp type ext2 (rw,nosuid,nodev,noatime)
none on /dev type tmpfs (rw,size=5M,mode=0755)
/dev/sde1 on /boot type ext2 (rw,errors=remount-ro)

and here is mdadm --device /dev/md0

/dev/md0:
Version : 00.90.01
Creation Time : Tue Apr 26 20:44:06 2005
Raid Level : raid5
Array Size : 725671872 (692.05 GiB 743.09 GB)
Device Size : 241890624 (230.68 GiB 247.70 GB)
Raid Devices : 4
Total Devices : 4
Preferred Minor : 0
Persistence : Superblock is persistent

Update Time : Sun Aug 21 09:37:31 2005
State : clean
Active Devices : 4
Working Devices : 4
Failed Devices : 0
Spare Devices : 0

Layout : left-symmetric
Chunk Size : 64K

UUID : b22bea6d:62339cd7:0ce83b75:4ac59414
Events : 0.4083507

Number Major Minor RaidDevice State
0 8 51 0 active sync /dev/sdd3
1 8 35 1 active sync /dev/sdc3
2 8 19 2 active sync /dev/sdb3
3 8 3 3 active sync /dev/sda3

Thanks for the help! Its a slow process, but I'm learning.

If udev is running then you won't have a problem. Udev creates /dev and everything in it on a temporary filesystem that will be deleted when you shut down and be recreated the next time you boot.

Feel free to reboot at your leasure.

I did the same thing, sudo rm -rf /opt/chroot and ended up in the same state with an empty /dev. I found the last message "Feel free to reboot at your leasure" to be pretty ominous, but I did it anyway and everything in /dev was recreated.

Thanks!

It seems I was still subscribed to this question...after reading the last post I realized I forgot to thank you for the help.

After rebooting, everything worked perfectly. (and still does!)

A few years late but: Thanks!

When you start the system, the "udev" facility examines the hardware configuration (according to established rules) and dynamically creates the entries in /dev.

"chroot" isn't the same thing.

It's a very nice feature. When you emptied the contents of /dev you of-course wiped out all those important entries, but it was "udev to the rescue!" You didn't have to worry about laboriously re-creating them: all you had to do was reboot.