VM started with VBoxHeadless hangs occasionally

[Freek de Kruijf]

I have an openSUSE 11.1 host system using VirtualBox 3.1.6. I have an openSUSE 11.1 guest. I run a cron job in the night on the host which stops the guest and it waits for the guest to be completely stopped. The cron jobs does its other work for about 45 minutes. After that the cron job starts the VM using the command:

VBoxManage startvm ktmh01 -type vrdp >> 00to02.log 2>/dev/null

Occasionally after this the cron job waits indefinitely for this command to finish. I have to stop VBoxHeadless using

kill -9 <pid>

Lately this happened on June 1st and June 8th. After killing VBoxHeadless I can start the VM using the same command and it works perfectly. I will prepare the cron job to enable the possibility to make a coredump when the the VM hangs again. However

sudo echo -n 1 > /proc/sys/fs/suid_dumpable

does not work, not even in console mode, but i will find a way to have a 1 in /proc/sys/fs/suid_dumpable when the VM is started and do a

ulimit -c unlimited

before the VBoxManage command.

[Frank Mehnert]

Thanks for the core dump. So far I found no obvious problem and I fear that such a core dump is not enough to debug your problem. It might help if you could send me a core dump of the VBoxHeadless process together with a backtrace of the corresponding VBoxSVC process the next time you observe such a hang. But it is even possible that only a debug log (special build required) would show the real problem.

Out of curiosity: Is there any special reason why you still use VBox 3.1.6 and not VBox 3.1.8?

[Freek de Kruijf]

This evening I will move to 3.1.8. The rest remains the same.

I have no idea how to make the suggested backtrace.

[Frank Mehnert]

You forced the core dump of the VirtualBox process. Just do the same with VBoxSVC.

[Klaus Espenlaub]

Does this problem still exist?

[Frank Mehnert]

No response, closing. Before you intend to reopen the ticket, please check the latest release 3.2.12.