VirtualBox

Ticket #11032 (new defect)

Opened 19 months ago

Last modified 18 months ago

VBox 4.2 on Solaris 11/11 with bonding locks up network adapter

Reported by: norchem Owned by:
Priority: critical Component: other
Version: VirtualBox 4.2.0 Keywords: bonding vnic solaris 4.2
Cc: Guest type: all
Host type: Solaris

Description

I am running Solaris 11/11 SRU 11.

I have two onboard network adapters which I had bonded. Both e1000e drivers.

I understood that virtualbox would not work properly with bond0 without vnics created. So I created a vnic for each Guest with the guests mac. After aprox 2 days of operation, the entire host machine will no longer be network accessible. There are no system log messages about any issues and VBoxSvr does not have any information. This of course crashes any VBox that is using iSCSI harddrives.

I was able to reproduce this issue twice. Each time taking 2 to 3 days for the problem to manifest.

I have since unbonded the adapters and set net0 as the primary bridgeadapter for all VMs and let the vbox network driver create its own vnics over net0. This has been running stable for 2 weeks now.

I have an identical hardware box also running Solaris 11/11 SRU 11 with bonding turned on. This box ONLY serves up iSCSI targets and has not exhibited the same behavior.

Hopefully I can get a third box where I can try reproducing it on that without disrupting production.

Change History

comment:1 Changed 18 months ago by ramshankar

Could you please provide more information on your setup? I assume by bonding you mean link aggregation.

Also if it is a problem that manifests after 2 to 3 days, chances are, unless we can reproduce this issue locally, fixing this would not be a priority (especially when nobody else has reported it so far).

At the very least, we'd require VBox.log for the VMs in question, output of: ifconfig -a, uname -a.

Note: See TracTickets for help on using tickets.

www.oracle.com
ContactPrivacy policyTerms of Use