I/O errors seen on tmpfs

Bug #1180021 reported by Clint Byrum
12
This bug affects 2 people
Affects Status Importance Assigned to Milestone
linux (Ubuntu)
Confirmed
Medium
Unassigned

Bug Description

Occasionally when using diskimage-builder on a large system with a large TMPFS, we see this error:

qemu-img: error while compressing sector 2069504: Input/output error

The exact operation that is happening is that qemu-img is taking a .raw file that is on a tmpfs partition and converting it to a qcow2 image. However, sometimes we also see errors when we create a loopback device and mkfs onto it.

The exact steps to reproduce:

git clone https://github.com/stackforge/diskimage-builder.git
cd diskimage-builder
bin/disk-image-create -x -a i386 base

It does not happen every time, maybe 1 in 10. However I have seen it on both Ubuntu 12.10 systems and Ubuntu 13.04, all on amd64. Also many times the errors are accompanied by these:

May 13 16:09:57 clint-HP kernel: [975578.653245] Buffer I/O error on device loop2p1, logical block 1943681
May 13 16:09:57 clint-HP kernel: [975578.653251] Buffer I/O error on device loop2p1, logical block 1943682
May 13 16:09:57 clint-HP kernel: [975578.653256] Buffer I/O error on device loop2p1, logical block 1943683
May 13 16:09:57 clint-HP kernel: [975578.653261] Buffer I/O error on device loop2p1, logical block 1943684
May 13 16:09:57 clint-HP kernel: [975578.653266] Buffer I/O error on device loop2p1, logical block 1943685
May 13 16:09:57 clint-HP kernel: [975578.653271] Buffer I/O error on device loop2p1, logical block 1943686
May 13 16:09:57 clint-HP kernel: [975578.653275] Buffer I/O error on device loop2p1, logical block 1943687
May 13 16:09:57 clint-HP kernel: [975578.653283] Buffer I/O error on device loop2p1, logical block 1943680
May 13 16:09:57 clint-HP kernel: [975578.653288] Buffer I/O error on device loop2p1, logical block 1943681

ProblemType: Bug
DistroRelease: Ubuntu 13.04
Package: linux-image-3.8.0-19-generic 3.8.0-19.30
ProcVersionSignature: Ubuntu 3.8.0-19.29-generic 3.8.8
Uname: Linux 3.8.0-19-generic x86_64
ApportVersion: 2.9.2-0ubuntu8
Architecture: amd64
AudioDevicesInUse:
 USER PID ACCESS COMMAND
 /dev/snd/controlC1: clint 3245 F.... pulseaudio
 /dev/snd/controlC0: clint 3245 F.... pulseaudio
Date: Tue May 14 10:05:58 2013
HibernationDevice: RESUME=UUID=b753173e-b967-403e-b076-b1cfb5bc6e80
InstallationDate: Installed on 2013-01-14 (120 days ago)
InstallationMedia: Ubuntu 12.10 "Quantal Quetzal" - Release amd64+mac (20121017.3)
MachineType: Hewlett-Packard HP EliteBook Folio 9470m
MarkForUpload: True
ProcEnviron:
 TERM=xterm
 PATH=(custom, no user)
 XDG_RUNTIME_DIR=<set>
 LANG=en_US.UTF-8
 SHELL=/bin/bash
ProcFB: 0 inteldrmfb
ProcKernelCmdLine: BOOT_IMAGE=/vmlinuz-3.8.0-19-generic root=/dev/mapper/ubuntu-root ro quiet splash vt.handoff=7
RelatedPackageVersions:
 linux-restricted-modules-3.8.0-19-generic N/A
 linux-backports-modules-3.8.0-19-generic N/A
 linux-firmware 1.106
SourcePackage: linux
UpgradeStatus: Upgraded to raring on 2013-04-30 (13 days ago)
dmi.bios.date: 09/24/2012
dmi.bios.vendor: Hewlett-Packard
dmi.bios.version: 68IBD Ver. F.30
dmi.board.name: 18DF
dmi.board.vendor: Hewlett-Packard
dmi.board.version: KBC Version 62.13
dmi.chassis.asset.tag: CNU2439PZT
dmi.chassis.type: 10
dmi.chassis.vendor: Hewlett-Packard
dmi.modalias: dmi:bvnHewlett-Packard:bvr68IBDVer.F.30:bd09/24/2012:svnHewlett-Packard:pnHPEliteBookFolio9470m:pvrA1029D1103:rvnHewlett-Packard:rn18DF:rvrKBCVersion62.13:cvnHewlett-Packard:ct10:cvr:
dmi.product.name: HP EliteBook Folio 9470m
dmi.product.version: A1029D1103
dmi.sys.vendor: Hewlett-Packard

Revision history for this message
Clint Byrum (clint-fewbar) wrote :
Revision history for this message
Brad Figg (brad-figg) wrote : Status changed to Confirmed

This change was made by a bot.

Changed in linux (Ubuntu):
status: New → Confirmed
Revision history for this message
Joseph Salisbury (jsalisbury) wrote :

Can you see if this can also be reproduced on the latest mainline kernel, version 3.10-rc1:
http://kernel.ubuntu.com/~kernel-ppa/mainline/v3.10-rc1-saucy/

Also, to reproduce this, can you provide some additional details on how large a system is required? Does it require a minimum amount of RAM, num cpus, etc? Does TMPFS need to be a minimum size?

Thanks in advance

Changed in linux (Ubuntu):
importance: Undecided → Medium
tags: added: kernel-da-key
Revision history for this message
Clint Byrum (clint-fewbar) wrote : Re: [Bug 1180021] Re: I/O errors seen on tmpfs

Excerpts from Joseph Salisbury's message of 2013-05-14 18:46:30 UTC:
> Can you see if this can also be reproduced on the latest mainline kernel, version 3.10-rc1:
> http://kernel.ubuntu.com/~kernel-ppa/mainline/v3.10-rc1-saucy/
>
> Also, to reproduce this, can you provide some additional details on how
> large a system is required? Does it require a minimum amount of RAM,
> num cpus, etc? Does TMPFS need to be a minimum size?
>

I have seen this on a system with 16G of RAM, and one with 96G of RAM.
Any less than 4G of RAM will result in diskimage-builder not using tmpfs.

Will try mainline kernel ASAP.

Revision history for this message
Phillip Susi (psusi) wrote :

What is the size limit on the tmpfs and how are you creating the image? My guess is that you create a sparse image that would exceed the size limit if it were not sparse, then when you try to write to it, you end up hitting the size limit.

Revision history for this message
Robert Collins (lifeless) wrote :
To post a comment you must log in.
This report contains Public information  
Everyone can see this information.

Other bug subscribers

Remote bug watches

Bug watches keep track of this bug in other bug trackers.