Hi,

This Quickstart version looks really polished, thanks for the effort of keeping up-to-date. :)

I had some problems with it.

I discovered the hard way... after 2 days customizing the virtual machine, stop working and didn't restart anymore with a Kernel Panic message = Init not found. try passing init=bootargs. (initramfs).

I thought was bad luck or something I did (like upgrading Ubuntu kernel), but after try many different things, going back in snapshots, whatever, it still happened.

I see that happens even with the default Quickstart 0.9.1. , but then it would be strange that nobody else has filled an issue with this problem. It's quite easy to reproduce this kernel panic....I tested that upgrading the recommended updates, include an upgrade to last kernel version and that will make the Kernel Panic message after rebooting. But it's not the only way, without upgrading also will happen soon or later...

Steps to fix it:
* Using VirtualBox 4.0.4, Windows 7 as guest machine and you must have another linux virtual machine that is bootable and ok.
* Download the Quickstart torrent, import the appliance. (don't start it)
* In your "other" virtual machine, Settings -> add a Hard Disk -> Choose the virtual disk of the quickstart appliance "Quickstart 0.91-disk1.vmdk" and it will go probably to SATA port 1.
* Start the"other" vm, and check that in the tool administration "Disk utility" what device it is.. should be probably /dev/sdb1 and Not Mounted!
* run e2fsck in the terminal on that device:

quickstart@qs091:~$ sudo e2fsck -f -v /dev/sdb1
e2fsck 1.41.12 (17-May-2010)
/dev/sdb1: recovering journal
Clearing orphaned inode 260625 (uid=114, gid=123, mode=0100600, size=0)
Clearing orphaned inode 260624 (uid=114, gid=123, mode=0100600, size=0)
Clearing orphaned inode 260621 (uid=114, gid=123, mode=0100600, size=0)
Clearing orphaned inode 260620 (uid=114, gid=123, mode=0100600, size=0)
Clearing orphaned inode 260619 (uid=114, gid=123, mode=0100600, size=0)
Pass 1: Checking inodes, blocks, and sizes
Pass 2: Checking directory structure
Pass 3: Checking directory connectivity
Pass 4: Checking reference counts
Pass 5: Checking group summary information

/dev/sdb1: ***** FILE SYSTEM WAS MODIFIED *****

  161023 inodes used (16.07%)
      74 non-contiguous files (0.0%)
     207 non-contiguous directories (0.1%)
         # of inodes with ind/dind/tind blocks: 0/0/0
         Extent depth histogram: 137385/54
  835526 blocks used (20.85%)
       0 bad blocks
       1 large file

  116697 regular files
   17971 directories
      59 character device files
      26 block device files
       0 fifos
     486 links
   26257 symbolic links (23485 fast symbolic links)
       4 sockets
--------
  161500 files
quickstart@qs091:~$ 

After this, you can shut down, remove the Hard Disk from the "other" virtual machine and start your Quickstart vm, with no problems.. I hope.

It seems to me that it's somehow corrupted.. and the more stuff that you do in the vm, the more corrupted it gets until a moment where it won't reboot anymore.

PS. I'm lazy to fill other issue, but also I see that installing manually Guest Addons through the Virtualbox interface, it will make stop working shared folders.

thanks
David

Comments

MichaelCole’s picture

Hi Corbacho,

Thanks for the awesome steps to fix the problem. This happened to another person I was working with in Austin. We decided it was easier to rebuild than fix. It also seemed related to the recent kernel update.

They were not shutting down the VM properly. E.g. shutdown from Ubuntu, or ACPI shutdown from Virtualbox. Just powering off the machine can cause the file system corruption you are describing.

Also, there is a recommended option in Virtualbox Settings->Storage->SATA Controller->Use host I/O cache.

My VBox recommended that this be enabled to prevent a known file system corruption bug.

Cheers,

Mike

MichaelCole’s picture

Category: bug » support
Status: Active » Fixed

Re: manually installing vbox extensions over existing ones. Don't do that :-)

corbacho’s picture

Hi Michael, thanks for reply.

Today, after that fix I did, I test to upgrade to last kernel and many other upgrades and customizations, reboot several times... still works ok, not a problem.

My host I/O cache is off... I saw in the documentation that if is ON it caches the I/O traffic, improving performance but making higher the risk of data lost or corruption.

But now you say the opposite? Or didn't I understand it?

A suggestion: Leave Ubuntu with "notify only" updates, not updating automatically. I did this way so I can control to make snapshots before upgrading. I don't want surprises any more, even this upgrade worked correctly.

PS. I read the Quickstart handbook and I saw that you shouldn't do manually the Guest Addons..., so I didn't do it this time with this last test... hopefully last test :D

MichaelCole’s picture

Hi Corbacho,

Re: Host I/O cache - I got a warning message to turn it ON in my version of Virtualbox. YMMV. Hopefully this will be fixed in the next version of VBox.

Re: Update. This shouldn't be a problem if VBox isn't crashing :-) You're welcome to turn this off for you - especially if you think it's tied to instability. Probably best to leave it on for the default though. Security patches should be "opt out".

Cheers,

Mike

MichaelCole’s picture

Status: Fixed » Closed (fixed)
corbacho’s picture

I didn't have warnings. I have it off . I think soon they will release a new version, it has been 2 months from the last one.

After my last comment, machine crashed one more time (after rebooting host PC the next day). Now I have a stable vm for 2 days. If someone else has problems.. I will recommend:
* Follow the steps of my first post to start with a clean table.
* Leave everything by default (I changed to 2 cpus since I have an i7 CPU and seems to be part of the reason of crashing)
* Don't update kernel without back up and disable security automatic updates (this includes the kernel), so they are not made in the background.
* I wouldn't trust so much in snapshots, they seem to get corrupt too. "export appliance" to make backups just in case.

Cheers
D

MichaelCole’s picture

Component: Code » Documentation
Category: support » feature
Status: Closed (fixed) » Active

Note to document the recovery process...

MichaelCole’s picture

Status: Active » Fixed

Status: Fixed » Closed (fixed)

Automatically closed -- issue fixed for 2 weeks with no activity.