My networks are already in different vlans. My issue is that the synchronization of ram, while migrating VMs, seems to be happening over the 10.135 network. For me this means that if you migrate a vm between machines that are in different racks... The traffic ends up traversing 3 switches before reaching its destination. This totally defeats the purpose of the totally redundant 10gig network I have in place.
So if I understand correctly 'pvecm create' uses '/etc/hostname' and then uses /etc/hosts to resolve its IP? ---- Phil K. > On Jan 16, 2015, at 8:39 AM, Steffen Wagner <[email protected]> wrote: > > You should not use the same NIC for Cluster / Storage and VM communications. > Therefore you should use a different setup. > Consider high Network I/O on VMs.. this might break your Cluster > communication. > You could also split your 10Gig network with vlans... > > Regards, > Steffen > > -----Original Message----- From: Phil Kauffman > Sent: Friday, January 16, 2015 3:36 PM > To: [email protected] > Subject: Re: [PVE-User] Cluster creation using specific IP or interface > > > And what if my automations require the host name to resolve to the > 10.135.164.x address? > > ---- > Phil K. > > On Jan 15, 2015, at 10:57 PM, Dietmar Maurer <[email protected]> wrote: > >>> Issue: When using ‘pvecm create <clustername>’ it chooses to use the >>> 10.135.164.x address. >>> >>> Can anyone tell me how ‘pvecm create’ get the IP it is going to use? Can I >>> influence it in someway? >> >> >> Edit /etc/hosts so that your hostname resolves to an address in the 10gig >> network. > _______________________________________________ > pve-user mailing list > [email protected] > http://pve.proxmox.com/cgi-bin/mailman/listinfo/pve-user _______________________________________________ pve-user mailing list [email protected] http://pve.proxmox.com/cgi-bin/mailman/listinfo/pve-user
