VMware Cloud Community
tomking_chen
Contributor
Contributor

vSphere 5.0 Distributed Switch bug

We have identified a vSphere 5.0 Distributed Switch bug with VMWare support a while back. The bug present as following:

We initially created the vSphere 5 cluster with two hosts and created Distributed Switch for those two ESXi 5 hosts. Later on we added another host into the cluster and applied same vDS configuration onto the added host. Then after sometime we notice some VMs in the cluster will randomly lost network connectivity after vMotion. We have to disconnect and re-connect the VM's vNIC to restore the connectivity. After escalate to VMWare support, they reckon this is due to the fact vDS profile didn't apply properly onto the newly added host. They consider this is a bug and advise us to completely re-build our vDS with all three hosts... Due to the scale of the impact we still haven't got a chance to execute the change.

After vSphere 5.0 Update 1 release, I read through its release notes, but could not identify anything in relate to our issue. So my guess is the issue still haven't been addressed.

Not sure if anyone out there experience the similar issue. Maybe you already have a better fix instead of re-build vDS completely?

Reply
0 Kudos
3 Replies
Linjo
Leadership
Leadership

Could you post your SR-number?

// Linjo

Best regards, Linjo Please follow me on twitter: @viewgeek If you find this information useful, please award points for "correct" or "helpful".
Reply
0 Kudos
tomking_chen
Contributor
Contributor

Hi, it was logged March this year. SR 12154834503

Reply
0 Kudos
evgf
Contributor
Contributor

Hello,

I experience a similar issue with a Distributed Switch with VLANs in our multi-host ESXi 5 environments.

The virtual machines from time to time experience network connectivity problems (sometimes only with a particular host or even the default gateway).

After disconnecting the VM from the dvSwitch and connecting it again the problems appear to be resolved, at least for some time.

If this is a known issue and there is a workaround, please share.

Kind Regards,

Evgeny

Reply
0 Kudos