Change VM - 'Warning: dracut-initqueue timeout' at startup

[GoodBytes]

Hello everyone,

I have CentOS 7 installed with VMWare, and I would like to change the VM to VirtualBox. I have done so, however CentOS7 refuses to boot successfully.
Clean install of the OS is not an option as it contains days of works setting it up.

When I start my VM under Virtual Box, instead of booting successfully to Gnome as expected and configured, it fails with the following error:

dracut-initqueue[259]: Warning: dracut-initqueue timeout
Warning: /dev/centos/root does not exist
Warning: /dev/centos/swap does not exist
Warning: /dev/mapper/centos-root does not exist

To be more exact:

boot.PNG (34.17 KiB) Viewed 177026 times

Here is what I tried:

1. Boot from CentOS 7 installation ISO disk
2. Hit Escape to get to the "Boot:" command, and executed: "linux rescue" command
3. Wait until it loads, and then pick Option 1 (Continue) to get to the shell.

   

   s1.PNG (27.16 KiB) Viewed 177026 times

4. Executed the following set of commands

   Code: Select all

   mount --bind /proc /mnt/sysimage/proc
   mount --bind /dev /mnt/sysimage/dev
   mount --bind /sys /mnt/sysimage/sys
   chroot /mnt/sysimage
   dracut -f 
   exit
   cd /
   sync
   telinit 6

   This sadly had no affect.
   
   It must be noted that everything seems to be executed successfully, beside "dracut -f" where I am getting:

   Kernel version 3.10.0-514.e17.x86_64 has no module directory /lib/modules/3.10.0-514.e17x86_64
   Failed to install module vmxnet3

My assumption is that 'vmxnet3' is a VMWare driver and that is screwing it up with the startup. Is this correct? And, if that is the case how do I remove it?

[TrevorH]

Did you try booting the "rescue" kernel from your installed system? Your problem is that the initramfs file for your current kernel only contains suport for the hardware that was present on your old VM. There should also be a "rescue" kernel listed in grub that was installed by the original installer and that should have support for most if not all hardware. You can boot that and then use dracut -f to rebuild the initramfs for the newest kernel and if you do that it will build a new initramfs containing support for the current VM "hardware".

[GoodBytes]

Hmm... sorry, you lost me a bit, how do I boot to rescue?

This is what I have:

boot_menu.PNG (25.24 KiB) Viewed 176981 times

I tried doing Ctrl+e on the selected item above (default) adding "rescue" (without quotes, of course), and run it (Ctrl+x), but I just get that it can't find rescue. I think I am not understanding you.

[TrevorH]

It's the bottom kernel on your list there - the one with "rescue" in its name Scroll down to that, hit enter.

[GoodBytes]

Wow.. I was indeed tired for not catching this.

Ok, interesting results:
1- Ran it first time, did something and quickly rebooted to the boot selection screen. Starting the OS normally, lead to the same issue.
2- Decided to try again to see if I can catch an error or something, but this time it did some what it looked like a disk scan (but correct me if I am wrong), thinking it would now work, I waited out (should have taken a screenshot for documentation purposes.... now I regret it). It rebooted, and starting the OS normally, once again lead to the same issue.
3- Ran it a third time to see what the scan was about, thinking it would redo it... but nope, now it worked! Logged in successfully, restarted normally, and now the normal option boots the OS perfectly fine.

Interesting set of event I wished I paid closer attention to know what exactly what was going on for self educational purposes. :/

But now everything works!

Thank you so much!
I really appreciate it!

[GoodBytes]

So, this is interesting. I am trying to move the VM from VMWare to Hyper-V (under Windows 10 if that matters) now, and I am getting different experience then above.
Basically, rescue boot option works, it successfully loads CentOS 7, but normal boot, it gets jam at:

Code: Select all

[    OK    ] Reached target Basic System

for a while, and then dracut timeout errors ("Warning: dracut-initqueue timeout - starting timeout script" appears a bunch of time), and then get and gets stuck (waited over 20min):

Code: Select all

[   OK   ] Started dracut initqueue hook.
[   OK   ] Reached target Remote File System (Pre).
[   OK   ] Reached target Remote File System.
[****     ] A start job is runnin for dev-mapp...oot.device (15min 40sec / no limit)

It must be noted, that when I boot in recovery mode. If I open a Terminal in Gnome, and type sudo dracut -f
I get the same error with dracut mentioned on my previous post "Failed to install module vmxnet3".

Any suggestions?

The reason why I want to move it to Hyper-V, is to have it running on the back with the system as it startups, and just SSH to it (I got this part working in a clean install of CentOS 7. But I need to port/use the setup and configured one.)

[TrevorH]

When you boot the rescue kernel it will use a different version number so you need to be explicit to dracut about which file to write and for what version. Don't just run `dracut -f` as it will build for the _running_ kernel version.

[GoodBytes]

Hmm.. I think I am not understanding the documentation of dracut.
I tried, under rescue mode > terminal:

Code: Select all

sudo dracut kernel-3.10.0-514.26.2.e17.x86_64.img -f

That didn't help.
I also tried

Code: Select all

sudo dracut kernel-3.10.0-514.26.2.e17.x86_64.img -f --omit-drivers "vmxnet3"

and

Code: Select all

sudo dracut kernel-3.10.0-514.26.2.e17.x86_64.img -o "vmxnet3"

And in all cases, dracut returns:

Kernel version 3.10.0-123.e17.x86_64 has no module directory /lib/modules/3.10.0-123e17.x86_64
dracut: Failed to install module wmxnet3

and of course, doesn't help when booting in normal mode.

If you wonder:
Doing:

Code: Select all

rpm -qa kernel

I get:

kernel-3.10.0-514.26.2.e17.x86_64
kernel-3.10.0-327.28.3.e17.x86_64
kernel-3.10.0-327.36.3.e17.x86_64
kernel-3.10.0-327.18.2.e17.x86_64
kernel-3.10.0-514.21.1.e17.x86_64

From the dracut error, it seems to ignore the image name, so I think that is my problem, but I can't see in the doc what I am doing wrong. I am definitely missing an argument it feels.

[hunter86_bg]

I usually use an installation DVD.
You have to boot the DVD, then select to rescue/repair (or whatever was called).
Then it will ask you whether you want to mount your file systems.
Once the file systems are mounted use

Code: Select all

mount -o rbind /proc /your-mounted-root-dpartition/proc

Repeat for '/sys' and '/dev' too.
Verify that your boot partition is mounted, if not mount it under your-mounted-root.
Then use

Code: Select all

chroot /your-mounted-root-partition

Last rebuild the initramfs, exit, dismount and reboot again.

It never failed me down.

[TrevorH]

All of the things you tried will not work. You need something like dracut -f /boot/initramfs-3.10.0-514.26.2.e17.x86_64.img 3.10.0-514.26.2.e17.x86_64

Pretty sure the message about vmxnet3 is a warning only and will not affect the outcome. That module is specific to a VMWare VM and won't be present anywhere else.