Hi there,
When i'm switching all my nodes to kernel 3.10.0-16-pve i've got the following messages appear in the dmesg :
UDP: bad checksum. From 172.22.3.103:5404 to 239.192.158.132:5405 ulen 83
I remember few month ago that i already tried to switch to the kernel 3.10 but when the UDP bad checksum messages appear all my nodes in the clusters were isolated from each other and the corrosync was down but now all is working no matter if the message was there or not.
I use openvswitch in my setup and the cluster management interface was an internal port connected to a bridge and the bridge was physically connected through a bond.
The messages are presents on my test and prod setup and there are completely different hardware configurations.
I've already tried to disable checksum and offload on the interfaces.
When i'm switching all my nodes to kernel 3.10.0-16-pve i've got the following messages appear in the dmesg :
UDP: bad checksum. From 172.22.3.103:5404 to 239.192.158.132:5405 ulen 83
I remember few month ago that i already tried to switch to the kernel 3.10 but when the UDP bad checksum messages appear all my nodes in the clusters were isolated from each other and the corrosync was down but now all is working no matter if the message was there or not.
I use openvswitch in my setup and the cluster management interface was an internal port connected to a bridge and the bridge was physically connected through a bond.
The messages are presents on my test and prod setup and there are completely different hardware configurations.
I've already tried to disable checksum and offload on the interfaces.