[PVE-User] About GRE tunnels and the new OVS bridging

Alexandre DERUMIER aderumier at odiso.com
Tue Apr 29 14:29:50 CEST 2014

>>No even with vxlan same disaster 

Are you sure that it's not a packet fragmentation problem ?

mtu should be increase on physical switchs to handle the overhead of vxlan (I think with ipsec too)


----- Mail original ----- 

De: "diaolin" <diaolin at diaolin.com> 
À: "Alexandre DERUMIER" <aderumier at odiso.com> 
Cc: "pve-user pve.proxmox.com" <pve-user at pve.proxmox.com> 
Envoyé: Lundi 28 Avril 2014 12:17:07 
Objet: Re: [PVE-User] About GRE tunnels and the new OVS bridging 

No even with vxlan same disaster 


Alexandre DERUMIER <aderumier at odiso.com> ha scritto: 

>Hi Diaolin, 
>maybe ovs vxlan tunnel could help ? maybe less overhead than a gre tunnel 
>----- Mail original ----- 
>De: "Diaolin" <diaolin at diaolin.com> 
>À: "pve-user pve.proxmox.com" <pve-user at pve.proxmox.com> 
>Envoyé: Mardi 22 Avril 2014 09:15:04 
>Objet: [PVE-User] About GRE tunnels and the new OVS bridging 
>As i said to Dietmar i tested it intensively for 
>adding a patch to adopt it in the PVE interface. 
>- very slow if the network is Gigabit (max 350Mbit) 
>- the transport has no encryption 
>I think that the better solution is to use a VLAN and 
>use the entire bandwidth (if needed). 
>I probed with openvpn crypted and uncrypted and i had 
>a better result (more than 50% of the net bandwidth) 
>This happens due to the use of bridging because if i use 
>(as route) a real interface and not a vmbr the result 
>are very good (OVS or OpenVPN). 
>And despite i didn't write the patch... 
>Sorry but if you intend to use a transport create one with 
>openvpn (even in MESH mode) that works better and is encrypted 
>S’à destacà l’ultima föia dal bósch nét 
>crodàda l’ei, solàgna, ‘n mèzz ai sàssi 
>e ‘ntant fis-ciava ‘n zìfol de oseleti 
>a tegnìr vìo ‘l pensér che vèn matìna 
>pve-user mailing list 
>pve-user at pve.proxmox.com 

More information about the pve-user mailing list