id summary reporter owner description type status component version resolution keywords cc guest host 13022 I/O errors reported by kernel for device backed by VDI Chris Murphy "/dev/sda backed by a VDI /dev/sdb backed by a vmdk, created with the command: sudo VBoxManage internalcommands createrawvmdk -filename /Users/chris/VirtualBox\ VMs/Fedora\ Rawhide/320g.vmdk -rawdisk /dev/disk2 /media/sf_chris/ is a Shared Folder whose source is HFS+ During sustained copying from /media/sf_chris and writing to /dev/sdb, the kernel reports spurius read and write errors from /dev/sda which is essentially inactive in this process. Long file copies (10 minutes) eventually result in segfaults, the system will no longer respond to commands and I can't even get kernel messages out of the VM to see what's gone wrong other than it looks like the I/O errors are effectively corrupting the OS as it's running. [ 222.202808] rawhide.localdomain kernel: sd 0:0:0:0: [sda] Unhandled error code [ 222.203003] rawhide.localdomain kernel: sd 0:0:0:0: [sda] [ 222.203287] rawhide.localdomain kernel: Result: hostbyte=DID_OK driverbyte=DRIVER_TIMEOUT [ 222.203407] rawhide.localdomain kernel: sd 0:0:0:0: [sda] CDB: [ 222.203528] rawhide.localdomain kernel: Write(10): 2a 00 00 58 a1 c8 00 02 00 00 [ 222.203644] rawhide.localdomain kernel: end_request: I/O error, dev sda, sector 5808584 [ 222.203746] rawhide.localdomain kernel: BTRFS: bdev /dev/sda3 errs: wr 1, rd 0, flush 0, corrupt 0, gen 0 [ 273.598086] rawhide.localdomain kernel: sd 0:0:0:0: [sda] Unhandled error code [ 273.598179] rawhide.localdomain kernel: sd 0:0:0:0: [sda] [ 273.598255] rawhide.localdomain kernel: Result: hostbyte=DID_OK driverbyte=DRIVER_TIMEOUT [ 273.598341] rawhide.localdomain kernel: sd 0:0:0:0: [sda] CDB: [ 273.598437] rawhide.localdomain kernel: Read(10): 28 00 01 17 73 40 00 00 20 00 [ 273.598533] rawhide.localdomain kernel: end_request: I/O error, dev sda, sector 18314048 [ 273.598618] rawhide.localdomain kernel: BTRFS: bdev /dev/sda3 errs: wr 1, rd 1, flush 0, corrupt 0, gen 0 The host kernel messages report no IO related errors. But does report other things related to VirtualBox at the time the problem occurs: May 8 23:04:16 ming kernel[0]: VBoxDrv: host_vmxoff -> vmx_use_count=0 May 8 23:04:44 ming kernel[0]: VirtualBox (map: 0xffffff8050dfec18) triggered DYLD shared region unnest for map: 0xffffff8050dfec18, region 0x7fff8ba00000->0x7fff8bc00000. While not abnormal for debuggers, this increases system memory footprint until the target exits. May 8 23:04:48 ming kernel[0]: VirtualBoxVM (map: 0xffffff803cc1b138) triggered DYLD shared region unnest for map: 0xffffff803cc1b138, region 0x7fff8ba00000->0x7fff8bc00000. While not abnormal for debuggers, this increases system memory footprint until the target exits. May 8 23:04:48 ming kernel[0]: VBoxDrv: host_vmxon -> vmx_use_count=1 rc=0 May 8 23:04:48 ming kernel[0]: VBoxDrv: host_vmxoff -> vmx_use_count=0 May 8 23:04:48 ming kernel[0]: VBoxDrv: host_vmxon -> vmx_use_count=1 rc=0 May 8 23:13:46 ming kernel[0]: VBoxDrv: host_vmxoff -> vmx_use_count=0 The vboxlog reports a lot of issues related to AHCI: 00:43:34.854982 AHCI#0P0: Cancelled task 7 00:43:34.855589 AHCI#0: Port 0 reset 00:43:50.165909 AHCI#0P0: Cancelled task 31 00:43:50.165928 AHCI#0: Port 0 reset 00:44:05.474666 AHCI#0P0: Cancelled task 31 00:44:05.474685 AHCI#0: Port 0 reset 00:44:35.781838 AHCI#0P0: Cancelled task 31 00:44:35.782117 AHCI#0: Port 0 reset 00:44:43.107712 AsyncCompletion: Task 0x00000131801a40 completed after 98 seconds 00:44:43.107728 AsyncCompletion: Task 0x00000131801ac0 completed after 98 seconds 00:44:43.107734 AsyncCompletion: Task 0x00000131801a80 completed after 98 seconds 00:44:43.108556 AsyncCompletion: Task 0x00000131801b40 completed after 98 seconds 00:44:43.108572 AsyncCompletion: Task 0x00000131801b00 completed after 98 seconds 00:44:43.108605 AsyncCompletion: Task 0x00000131801b80 completed after 98 seconds 00:44:43.148090 AHCI#0P0: Flush request was active for 98 seconds 00:44:43.148101 AHCI#0P0: Canceled flush returned rc=VINF_SUCCESS 00:44:43.148108 AsyncCompletion: Task 0x00000131801bc0 completed after 98 seconds 00:44:43.473168 AHCI#0P0: Flush request was active for 68 seconds 00:44:43.473183 AHCI#0P0: Canceled flush returned rc=VINF_SUCCESS 00:44:44.688476 AHCI#0P0: Flush request was active for 54 seconds 00:44:44.688495 AHCI#0P0: Canceled flush returned rc=VINF_SUCCESS 00:44:44.688838 AHCI#0P0: Flush request was active for 38 seconds 00:44:44.688850 AHCI#0P0: Canceled flush returned rc=VINF_SUCCESS Since this is reporting the problem is with sda and not sdb, I don't think it's the physical external drive causing the problem, it's either a guest additions bug or how AHCI is implemented in VBox and interacting with Linux. kernel-3.15.0-0.rc4.git0.1.fc21.x86_64" defect new other VirtualBox 4.3.10 Linux Mac OS X