Hi All,
I have spend a few hours researching bonds+bridges in relation to Proxmox, Debian Lenny and OpenVZ with the intent of creating a separate bridge for each vm in order to run Munin for tracking network usage on the vms individually. No VLANs have been used in my testing. Some things are just not clear to me.
I'll keep this really simple and ask my questions in point form. The main gist of things is questioning the ability to manipulate network configuration without rebooting host or vms and to try and understand the correct build up methods for bonding+bridging.
1. When adding bonded and bridging devices via the GUI one must reboot for it to take effect. It is possible to add them manually via /etc/network/interfaces and restart the networking but I have found that there is all kinds of wonkyness that occurs (eg cant ping vms from host or network and vice versa). Sometimes restarting all the VMs will correct this. Does someone have a step by step procedure for doing this manually without needing to reboot the host or the vms?
2. Like the question above, when adding multiple bridges to a single bond (I haven't tried with multiple bonds), it seems that only a reboot will initialize them. When adding/changing the bridge to the vm for use, the vm must be restarted for the configuration to take effect. Although it is possible to manually add multiple bridges to a single bond (tested with same subnet) via /etc/network/interfaces. However in the GUI (and manually) many times there is a complaint about not being able to use a bond that has been assigned already to the first bridge. A reboot will fix this complaint but another issue is that there is only routing to the first bridge in the list and the other vms behind the other bridges become isolated. Is this a bug or correct functionality?
3. Do I have to put each vm/bridge into separate VLANs for this to work properly?
4. How can I further investigate these issues?
Im sure I have more questions but thats it for now.
I have spend a few hours researching bonds+bridges in relation to Proxmox, Debian Lenny and OpenVZ with the intent of creating a separate bridge for each vm in order to run Munin for tracking network usage on the vms individually. No VLANs have been used in my testing. Some things are just not clear to me.
I'll keep this really simple and ask my questions in point form. The main gist of things is questioning the ability to manipulate network configuration without rebooting host or vms and to try and understand the correct build up methods for bonding+bridging.
1. When adding bonded and bridging devices via the GUI one must reboot for it to take effect. It is possible to add them manually via /etc/network/interfaces and restart the networking but I have found that there is all kinds of wonkyness that occurs (eg cant ping vms from host or network and vice versa). Sometimes restarting all the VMs will correct this. Does someone have a step by step procedure for doing this manually without needing to reboot the host or the vms?
2. Like the question above, when adding multiple bridges to a single bond (I haven't tried with multiple bonds), it seems that only a reboot will initialize them. When adding/changing the bridge to the vm for use, the vm must be restarted for the configuration to take effect. Although it is possible to manually add multiple bridges to a single bond (tested with same subnet) via /etc/network/interfaces. However in the GUI (and manually) many times there is a complaint about not being able to use a bond that has been assigned already to the first bridge. A reboot will fix this complaint but another issue is that there is only routing to the first bridge in the list and the other vms behind the other bridges become isolated. Is this a bug or correct functionality?
3. Do I have to put each vm/bridge into separate VLANs for this to work properly?
4. How can I further investigate these issues?
Im sure I have more questions but thats it for now.