Virtual disk automatic resize/growth failing.

[George R. Goffe]

Hi,

When the vitual hdd is defined during VM creation, I selected a dynamically allocated VDI. As far as I can tell, the size selected will be the maximum size of the hdd on the host and that less space will be allocated initially and extended as needed during guest operation. Correct?

My situation is a VM with a dynamically allocated hdd of 384GB. Right now it's at 45GB or so but is trying to extend the size. The guest is reporting disk problems and out of space problems and has now changed one partition to read only. What I think is happening is that the guest has a code path which Virtual Box intercepts, does some magic to add space, restart the operation. This is apparently NOT happening. The path through Virtual Box may or may not be trying to handle this situation but then appears to let the error(s) continue through the guest OS (5.12.0-0.rc5.180.fc35.x86_64).

Here's what the host thinks of the partition where the VDI is located:

Filesystem                         1K-blocks        Used   Available  Really_Available Use% Mount_Point
/dev/sda8                         1735405656   485065980  1250569508        1250339676  28% /export/home

The guest thinks it's VDI is:

Filesystem     1K-blocks     Used Available Use% Mounted on
/dev/sda2        1992552   164592   1706720   9% /
/dev/sda7       65478188 14001584  48104780  23% /usr
/dev/sda1        1992552   149816   1721496   9% /boot
/dev/sda8      184666464 19495796 155717356  12% /export/home
/dev/sda5       65478188   751092  61355272   2% /var
/dev/sda6       65478188   936844  61169520   2% /opt

Which is considerably less than 384GB.

Additionally, the guest is throwing numerous messages like these:

[   23.180382] 06:32:17.603879 main     VBoxService 6.1.18_Fedora r142142 (verbosity: 0) linux.amd64 (Mar  2 2021 00:00:00
) release log
               06:32:
[   23.184678] 06:32:17.610514 main     OS Product: Linux
[   23.189518] 06:32:17.614980 main     OS Release: 5.12.0-0.rc5.180.fc35.x86_64
[   23.192546] 06:32:17.617869 main     OS Version: #1 SMP Mon Mar 29 15:30:36 UTC 2021
[   23.197236] 06:32:17.620938 main     Executable: /usr/sbin/VBoxService
               06:32:17.620942 main     Process ID: 591
               06:32:17.620944 main     Pa
[   23.202542] 06:32:17.628117 main     6.1.18_Fedora r142142 started. Verbose level = 0
[   23.208141] 06:32:17.633605 main     vbglR3GuestCtrlDetectPeekGetCancelSupport: Supported (#1)
[   40.487920] e1000: enp0s3 NIC Link is Up 1000 Mbps Full Duplex, Flow Control: RX
[   40.488885] IPv6: ADDRCONF(NETDEV_CHANGE): enp0s3: link becomes ready
[ 7841.824091] systemd[1]: systemd v248-1.fc35 running in system mode. (+PAM +AUDIT +SELINUX -APPARMOR +IMA +SMACK +SECCOM
P +GCRYPT +GNUTLS +OPENSSL +ACL +BLKID +CURL +ELFUTILS +FIDO2 +IDN2 -IDN +IPTC +KMOD +LIBCRYPTSETUP +LIBFDISK +PCRE2 +PWQU
ALITY +P11KIT +QRENCODE +BZIP2 +LZ4 +XZ +ZLIB +ZSTD +XKBCOMMON +UTMP +SYSVINIT default-hierarchy=unified)
[ 7841.825374] systemd[1]: Detected virtualization oracle.
[ 7841.825631] systemd[1]: Detected architecture x86-64.
[ 7872.660143] ata1.00: exception Emask 0x0 SAct 0x1ffff800 SErr 0x0 action 0x6 frozen
[ 7872.661528] ata1.00: failed command: WRITE FPDMA QUEUED
[ 7872.662445] ata1.00: cmd 61/00:58:c8:89:2b/0a:00:12:00:00/40 tag 11 ncq dma 1310720 ou
                        res 40/00:00:00:00:00/00:00:00:00:00/00 Emask 0x4 (timeout)
[ 7872.664910] ata1.00: status: { DRDY }
[ 7872.665671] ata1.00: failed command: WRITE FPDMA QUEUED
[ 7872.666963] ata1.00: cmd 61/00:60:c8:93:2b/0a:00:12:00:00/40 tag 12 ncq dma 1310720 ou
                        res 40/00:00:00:00:00/00:00:00:00:00/00 Emask 0x4 (timeout)

and:

[ 7872.721207] ata1.00: status: { DRDY }
[ 7872.721577] ata1.00: failed command: READ FPDMA QUEUED
[ 7872.722023] ata1.00: cmd 60/08:e0:10:7a:0d/00:00:17:00:00/40 tag 28 ncq dma 4096 in
                        res 40/00:00:00:00:00/00:00:00:00:00/00 Emask 0x4 (timeout)
[ 7872.723230] ata1.00: status: { DRDY }
[ 7872.723607] ata1: hard resetting link
[ 7873.036871] ata1: SATA link up 3.0 Gbps (SStatus 123 SControl 300)
[ 7873.038028] ata1.00: configured for UDMA/133
[ 7873.038418] ata1.00: device reported invalid CHS sector 0
[ 7873.038850] ata1.00: device reported invalid CHS sector 0
[ 7873.039295] ata1.00: device reported invalid CHS sector 0
[ 7873.039725] ata1.00: device reported invalid CHS sector 0
[ 7873.040250] ata1.00: device reported invalid CHS sector 0
[ 7873.040691] ata1.00: device reported invalid CHS sector 0
[ 7873.041133] ata1.00: device reported invalid CHS sector 0
[ 7873.041646] ata1.00: device reported invalid CHS sector 0

and:

[49175.522760] EXT4-fs (sda8): failed to convert unwritten extents to written extents -- potential data loss!  (inode 2621957, error -30)
[49175.526567] EXT4-fs (sda8): failed to convert unwritten extents to written extents -- potential data loss!  (inode 2621958, error -30)
[49175.530010] EXT4-fs (sda8): failed to convert unwritten extents to written extents -- potential data loss!  (inode 2621961, error -30)

Of course there are more of all of these messages available in a file. I'm not sure how I would attach a file to this mailing list though.

Can someone take a look at all this and render some assistance please?

Best regards,

George...

VirtualBox Graphical User Interface

Version 6.1.19 r143581 (Qt5.6.1)

[Klaus Espenlaub]

You provided a lot of useful information, but a key item is missing: VBox.log (if you use default layout then it's in ~/VirtualBox VMs/<vmname>/Logs), showing the affected VM run.

The logging from the guest side points at I/O timeouts, which are outside the control of VirtualBox in the general case. The VBox.log file would allow to tell more what happened on the host side. It is odd that the AHCI emulation does not recover from these timeouts properly. Known to work unless the host's disk is unresponsive.

It is clear that you're not running out of disk space on the host (or in the VM). Your understanding of dynamic images is correct, they save on how much space is consumed on the host and grow up to the selected capacity.

[George R. Goffe]

Klaus,

Thank you for responding to this bug report, it's appreciated.

If you need any other info or want me to try other versions of VB, just let me know. I want to help.

Best regards and stay safe,

George...

[George R. Goffe]

Hi,

Is anyone working this ticket?

I just tried the latest 6.1.9 VB and it gives exit code 1 at startup.

What's going on guys?

George... er... obiewan... :-)

[aeichner]

Your VDI is stored on a BTRFS filesystem which is known to cause problems, you can try moving the VM to a different filesystem or try to enable the "Use Host I/O Cache" setting for the VM in the storage settings.

[George R. Goffe]

Howdy,

Thank you for responding to this "bug" report, your help, time, effort IS APPRECIATED!

Best regards and STAY SAFE,

George... er... obiewan... :-)