qemu-system-ppc : file systems are not shutting down clean

Bug #1321028 reported by john.d@servergy.com
6
This bug affects 1 person
Affects Status Importance Assigned to Milestone
QEMU
New
Undecided
Unassigned
qemu (Ubuntu)
New
Undecided
Unassigned

Bug Description

Launching a VM that has been shutdown gracefully ( via init 0) typically requires fsck to run when it is started ;
This indications data integrity issues;

 The symptom can be seen by observing the fsck running when the VM restarted.

Install 14-04-LTS to a VM and the issue can be seen ;

(trusty)vnc@jade-rev4:/home2/qemu$ cat vm1/go.sh
mymac="52:54:00:12:34:10"
T=` echo $mymac | cut -d: -f5,6 | sed s/\://`
mytap="tap$T"

tunctl -d $mytap
tunctl -t $mytap

 /usr/bin/qemu-system-ppc -M ppce500 -nographic -kernel jade-kernel.bin \
  -initrd jade-initrd-2.0.bin -m 1G -enable-kvm \
  -drive file=jade-ubuntu-14.04.raw,if=virtio \
  -net nic,vlan=0,macaddr=$mymac \
  -net tap,vlan=0,ifname=$mytap,script=/usr/local/bin/qemu-ifup \
  -append "console=ttyS0 ssgyboot=break" \
   -no-shutdown -no-reboot -name `basename $fp`

ProblemType: Bug
DistroRelease: Ubuntu 14.04
Package: qemu-system-ppc 2.0.0~rc1+dfsg-0ubuntu3.1
ProcVersionSignature: Ubuntu 3.13.0-24.46-powerpc-e500mc 3.13.9
Uname: Linux 3.13.0-24-powerpc-e500mc ppc
ApportVersion: 2.14.1-0ubuntu3
Architecture: powerpc
Date: Mon May 19 17:01:14 2014
ProcEnviron:
 TERM=xterm
 PATH=(custom, no user)
 LANG=en_US.UTF-8
 SHELL=/bin/bash
SourcePackage: qemu
UpgradeStatus: Upgraded to trusty on 2014-04-29 (20 days ago)

Revision history for this message
john.d@servergy.com (john-d-0) wrote :
Revision history for this message
Serge Hallyn (serge-hallyn) wrote :

Can you explain what "ssgyboot=break" tells the kernel to do?

Could you upload the guest's /var/log/syslog after reboot?

Please show

echo $?

immediately after the qemu-system-ppc command has exited?

Changed in qemu (Ubuntu):
status: New → Incomplete
Revision history for this message
john.d@servergy.com (john-d-0) wrote :

1. ssgyboot-break has no effect in the VM kernel; It is only used by jade-initrd-2.0.bin ;

  kernel cmdline:

[ 0.000000] pcpu-alloc: s28800 r8192 d16256 u53248 alloc=13*4096
[ 0.000000] pcpu-alloc: [0] 0
[ 0.000000] Built 1 zonelists in Zone order, mobility grouping on. Total pag
es: 260608
[ 0.000000] Kernel command line: (host)/boot/vmlinux-3.13.0-24-powerpc-e500mc
 root=/dev/vda1 ro quiet splash vt.handoff=7

2: see attached file .

3. VM's do not exit from qemu-system-ppc ; see defect : 1317603 .

Revision history for this message
john.d@servergy.com (john-d-0) wrote :

/var/log/syslog attached as syslog.onboot when file system is dirty ;

Since the VM terminated ; the device should have been umounted ;

Revision history for this message
john.d@servergy.com (john-d-0) wrote :

I am thinking since the VM hasn't terminated due to defect 1317603 , qemu-system-ppc has to be terminated with kill <pid> which is not kill -9 ; I could see some inconsistencies with virtio device file not being sync'ed ; but the VM did unmount it;

Changed in qemu (Ubuntu):
status: Incomplete → New
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.