Enabling swap failed [solved]

[dfgrbac]

I recently upgraded from Mandriva 2005 Limited Edition to Mandriva 2007. On booting Mandriva 2007 I get the error message enabling swap failed - device or resource busy.

 

I found an older discussion on this swap problem that seems to indicate that it is really not a problem, but has something to do with logical volumes. If this is really true, the error message should be deleted from the boot process since it scares users. I have been using Linux systems since Slackware 0.97 and have never had a swap problem until now. Mandriva 2005 used logical volumes also and didn't have the problem. I did the suggestions in that discussion with the following results:

 

[dfgrbac@localhost ~]$ cat /etc/fstab
# /dev/VolGroup00/LogVol00, size=119013376, type=:  (VolGroup00)
/dev/VolGroup00/LogVol00 / ext3 defaults 1 1
# /dev/hda5, size=208782, type=131: Journalised FS: ext3 (extended)
/dev/hda5 /boot ext3 defaults 1 2
none /dev/pts devpts mode=0620 0 0
# /dev/VolGroup00/2, size=1572864, type=:  (VolGroup00)
/dev/VolGroup00/2 /mnt/VolGroup00/2 ext2 user,exec,rw,noauto 1 2
# fd: H1440
/dev/fd0 /mnt/floppy auto user,exec,rw,noauto 0 0
# /dev/hda1, size=9858177, type=11: FAT32 (primary)
/dev/hda1 /mnt/win_c vfat user,exec,rw,noauto,iocharset=iso8859-1,codepage=850,umask=0 0 0
# /dev/hda2, size=93337650, type=7: NTFS (primary)
/dev/hda2 /mnt/win_d ntfs user,exec,ro,noauto,iocharset=iso8859-1,umask=0 0 0
# /dev/hda3, size=8385930, type=12: Win98 FAT32, LBA-mapped (primary)
/dev/hda3 /mnt/win_e vfat user,exec,rw,noauto,iocharset=iso8859-1,codepage=850,umask=0 0 0
/dev/sda1 /mnt/zipext ext2 defaults 0 0
/dev/sda4 /mnt/zip vfat defaults 0 0
none /proc proc defaults 0 0
none /proc/bus/usb usbfs defaults 0 0
# /dev/VolGroup00/LogVol01, size=2031616, type=:  (VolGroup00)
/dev/VolGroup00/LogVol01 swap swap defaults 0 0
[dfgrbac@localhost ~]$

 

Swap is listed there.

 

[dfgrbac@localhost ~]$ free
		 total	   used	   free	 shared	buffers	 cached
Mem:		508128	 460656	  47472		  0	  34844	 174052
-/+ buffers/cache:	 251760	 256368
Swap:	  1015800		  0	1015800
[dfgrbac@localhost ~]$

 

And swap is listed again. Does this mean it is working okay and I should ignore the error?

[ianw1974]

I'd be tempted to do this:

 

swapoff /dev/VolGroup00/LogVol01

 

and then make it again:

 

mkswap /dev/VolGroup00/LogVol01

 

and see if that helps. Although, perhaps not do this if you have more than one distro on the machine, in case the other distro starts complaining. But the complaining in the other distro will probably be due to the settings in the /etc/fstab so you should be able to fix it easy enough by changing the /etc/fstab. If you only have one distro installed, go for it. It won't hurt.

[iphitus]

The swap is listed as available, so it doesnt need to be mkswapped. Even if you did that, it would have zero effect on the other distro.

 

From what you say "device or resource busy" it sounds like for some reason, the mandriva boot scripts are attempting to mount the swap *twice*. The first time, as you can see is successful, the second time errors out.

 

This is completely harmless.

 

Just remember, this isnt an official mandriva board, so you'll have more luck getting rid of the error by filing a bug at the mandriva bugzilla.

 

James

[dfgrbac]

It appears that the error message is itself an error - indeed. The Drak Control Center partition page does not show the green swap partition I created long ago. But when I click on the logical volume, the green swap space appears there in the size I created.

 

Also, swapoff and swapon works correctly from the root console window. So the boot message is indeed an error!

 

Thanks for the discussion. I feel better now.