VirtualBox frequently aborting with VERR_IO_CRC

[Andy2015]

Hi,

I'm using VirtualBox on Windows 7 running Lubuntu 15.04. I'm using it for running Docker as well as other Linux applications. When doing innocuous things (e.g. opening firefox) the VM will go into an Aborted state. The logs report the following:

Code: Select all

00:28:45.482137 AioMgr0-N: Request 0x0000000ca986b0 failed with rc=VERR_IO_CRC, migrating endpoint C:\Users\andy\VirtualBox VMs\Lubuntu 40GB\Lubuntu 40GB.vdi to failsafe manager.
00:28:45.729653 AHCI#0P0: Read at offset 2568564736 (131072 bytes left) returned rc=VERR_IO_CRC
00:28:45.969613 AHCI#0P0: Read at offset 2568564736 (131072 bytes left) returned rc=VERR_IO_CRC
00:28:46.110473 AHCI#0P0: Read at offset 2568564736 (131072 bytes left) returned rc=VERR_IO_CRC
00:28:46.188569 AHCI#0P0: Read at offset 2568564736 (131072 bytes left) returned rc=VERR_IO_CRC
00:28:46.264603 AHCI#0P0: Read at offset 2568564736 (131072 bytes left) returned rc=VERR_IO_CRC
00:28:46.330682 AHCI#0P0: Read at offset 2568564736 (131072 bytes left) returned rc=VERR_IO_CRC
00:28:46.377807 AHCI#0P0: Canceled read at offset 2604052480 (131072 bytes left) returned rc=VINF_SUCCESS

Versions:
Windows 7 SP1 64 bit with SSD hard drive
VirtualBox v4.3.28

Any suggestions on how to fix this?

Many Thanks

Andy

[mhanor]

Is your HDD healthy?
Anyway, you're hitting a know issue with the VirtualBox SATA controller. You can try to enable the Host I/O Cache option, in the VM settings -> Storage -> Controller SATA. If that doesn't work, attach the virtual disk to another type of controller.

[mpack]

Hmm. I'm not aware of that known issue. It looks like bad sectors on the host drive to me.

[mhanor]

https://www.virtualbox.org/ticket/13105

[mpack]

It would be useful if the OP posted a complete VM log, not an extract. Zip it before attaching here.

[Andy2015]

I've attached the full log in a zip file. On running chkdsk I have seen a few bad sectors that have now been fixed. Will run some more diagnostics on the drive itself.

[Andy2015]

I also ticked VM settings -> Storage -> Controller SATA -> Use Host I/O Cache. However the aborts are still happening.
@mhanor - would you suggest using a particular type of controller for the disk? e.g. SCSI?

Currently running firefox (which was previously working) from within the Lubuntu VM causes an Application Error popup to appear:
"The exception Breakpoint. A breakpoint has been reached. (0x80000003) occurred in the application at location 0xe08fcb52. Click on OK to terminate the program"

The last lines of the log file are as follows (full log file attached):

Code: Select all

19:54:42.754602 AHCI#0P0: Read at offset 2568564736 (131072 bytes left) returned rc=VERR_IO_CRC
19:54:42.804480 AHCI#0P0: Read at offset 2568564736 (131072 bytes left) returned rc=VERR_IO_CRC
19:54:42.811709 AHCI#0: Port 0 reset
19:54:43.163785 AHCI#0P0: Read at offset 2568564736 (131072 bytes left) returned rc=VERR_IO_CRC
19:54:43.179434 AHCI#0P0: Cancelled task 28
19:54:43.179532 
19:54:43.179534 !!Assertion Failed!!
19:54:43.179535 Expression: ASMAtomicReadU32(&pAhciPort->cTasksActive) > 0
19:54:43.179537 Location  : F:\tinderbox\win-4.3\src\VBox\Devices\Storage\DevAHCI.cpp(6069) bool __cdecl ahciTransferComplete(struct AHCIPort *,struct AHCIREQ *,int,bool)
19:54:43.179600 Inconsistent request counter
19:54:48.179533 AHCI#0: Port 0 reset

[mhanor]

You should run an offline full surface scan.
You can choose any other controller, but the guest OS might fail to boot. If that's the case, you'll have to add the right driver to the initrd of the guest OS.

[Andy2015]

chkdsk has found 700KB of bad sectors on the 114GB SSD. Is this anything to worry about?

previously we've attempted to repair this using chkdsk, but the bad sectors aren't going away. (I think it's been around 700KB for the last month or so, so does not appear to be increasing)

[mpack]

Yes, bad sectors on the drive are something to be concerned about. All pre-existing bad sectors should have been mapped and bypassed in the factory. The fact that you have new ones means that either you did something one-off which damaged the drive (e.g. if a laptop, you dropped it), or more seriously that the drive is spontaneously developing new bad sectors, which would be a prelude to drive failure.

When I had something similar happen, I replaced the drive. At the very least you should increase the frequency of your whole drive backups - and continue to monitor the bad sector count.

Oh, and go ahead and fix the current crop of bad sectors (I'm sure you know that you can't actually fix them - Windows just marks and avoids them in future).