Re: [one-users] Ranged network!
To improve this, we can extend the parameters to define the ranged network, e.g. with the starting ip address parameter. Or alternatively, define a fixed net using single leases and/or continuous ranges of leases. Which one is better? I think that RANGED networks are ultimately easier to manage than a FIXED network. I would like, for example, to create several small networks out of a larger block of addresses...for example, taking 10.10.10.0/24 and dividing it up into: - 10.10.10.10 - 10.10.10.30 - 10.10.10.31 - 10.10.10.100 - etc... It's certainly possible to do this using FIXED networks and generated leases, but it's somewhat less convenient and doesn't really lend itself to management via the web interface. This would require the ability to specify a base address for each network, as well as the the number of addresses to provision. So something like this: NAME=cloud TYPE=RANGED BRIDIGE=br619 NETWORK_BASE=10.10.10.10 NETWORK_SIZE=30 Ideally, I would like to be able to specify the network using any of the following: - An address mask: NETWORK_ADDRESS=10.10.10.0/24 - An address and mask as octets: NETWORK_ADDRESS=10.10.10.0/255.255.255.0 - A base and size: NETWORK_BASE=10.10.10.1 NETWORK_SIZE=253 (where the first address allocated is NETWORK_BASE, and the last address allocated is NETWORK_BASE+NETWORK_SIZE). It would also be nice to be able to reserve addresses on a ranged network. For example, we have a /24 dedicated to our OpenNebula cluster right now, except there are two legacy hosts sitting there from a previous virtualization experiment. It would be great if I could do this: NETWORK_ADDRESS=10.10.10.0/24 RESERVED=[10.10.10.1, 10.10.10.254, 10.10.10.253] ...or something like that. -- Lars Kellogg-Stedman l...@seas.harvard.edu Senior Technologist Harvard University SEAS Academic and Research Computing (ARC) ___ Users mailing list Users@lists.opennebula.org http://lists.opennebula.org/listinfo.cgi/users-opennebula.org
Re: [one-users] Ranged network!
Thanks for sharing!,... I've filled a ticket to track and implement this http://dev.opennebula.org/issues/602 Cheers Ruben On Fri, Apr 29, 2011 at 10:36 PM, Lars Kellogg-Stedman l...@seas.harvard.edu wrote: To improve this, we can extend the parameters to define the ranged network, e.g. with the starting ip address parameter. Or alternatively, define a fixed net using single leases and/or continuous ranges of leases. Which one is better? I think that RANGED networks are ultimately easier to manage than a FIXED network. I would like, for example, to create several small networks out of a larger block of addresses...for example, taking 10.10.10.0/24 and dividing it up into: - 10.10.10.10 - 10.10.10.30 - 10.10.10.31 - 10.10.10.100 - etc... It's certainly possible to do this using FIXED networks and generated leases, but it's somewhat less convenient and doesn't really lend itself to management via the web interface. This would require the ability to specify a base address for each network, as well as the the number of addresses to provision. So something like this: NAME=cloud TYPE=RANGED BRIDIGE=br619 NETWORK_BASE=10.10.10.10 NETWORK_SIZE=30 Ideally, I would like to be able to specify the network using any of the following: - An address mask: NETWORK_ADDRESS=10.10.10.0/24 - An address and mask as octets: NETWORK_ADDRESS=10.10.10.0/255.255.255.0 - A base and size: NETWORK_BASE=10.10.10.1 NETWORK_SIZE=253 (where the first address allocated is NETWORK_BASE, and the last address allocated is NETWORK_BASE+NETWORK_SIZE). It would also be nice to be able to reserve addresses on a ranged network. For example, we have a /24 dedicated to our OpenNebula cluster right now, except there are two legacy hosts sitting there from a previous virtualization experiment. It would be great if I could do this: NETWORK_ADDRESS=10.10.10.0/24 RESERVED=[10.10.10.1, 10.10.10.254, 10.10.10.253] ...or something like that. -- Lars Kellogg-Stedman l...@seas.harvard.edu Senior Technologist Harvard University SEAS Academic and Research Computing (ARC) ___ Users mailing list Users@lists.opennebula.org http://lists.opennebula.org/listinfo.cgi/users-opennebula.org -- Dr. Ruben Santiago Montero Associate Professor (Profesor Titular), Complutense University of Madrid URL: http://dsa-research.org/doku.php?id=people:ruben Weblog: http://blog.dsa-research.org/?author=7 ___ Users mailing list Users@lists.opennebula.org http://lists.opennebula.org/listinfo.cgi/users-opennebula.org
Re: [one-users] Ranged network!
we have setup up a ranged virtual network for our ONE-2.2 VMs but the gateway IP address keeps being assigned to VMs. Ahmad, I just asked a very similar question on the list. I agree that the existing documentation isn't particularly clear. The only response I received was a suggestion to use FIXED networks rather than RANGED networks, and then to programmatically generate the list of leases. So our network configuration looks like this: BRIDGE=br619 NAME=Cloud Private TYPE=FIXED Followed by the output of the following shell script: for (( ip=20; ip255; ip++ )); do printf LEASES = [IP=10.243.19.%d, MAC=5E:A5:%02X:%02X:%02X:%02X]\n $ip 10 243 19 $ip done This generates output that looks like this: LEASES = [IP=10.243.19.20, MAC=5E:A5:0A:F3:13:14] LEASES = [IP=10.243.19.21, MAC=5E:A5:0A:F3:13:15] [...] LEASES = [IP=10.243.19.254, MAC=5E:A5:0A:F3:13:FE] In this case, we're using the MAC prefix model, so we're able to pre-generate this list. This has solved the problem of OpenNebula handing out the router address to virtual machines, at the cost of a small amount of complexity. -- Lars Kellogg-Stedman l...@seas.harvard.edu Senior Technologist Harvard University SEAS Academic and Research Computing (ARC) ___ Users mailing list Users@lists.opennebula.org http://lists.opennebula.org/listinfo.cgi/users-opennebula.org
Re: [one-users] Ranged network!
Hi, Please also note that is very common to put the gateway in a VM also as part of the network. You can request an specific IP for the gateway VM in the template (NIC=[ NETWORK_ID = 3, IP = 192.168.0.1]) Probably, we should make more clear the documentation to remark that you are defining the range using a network address and not a starting host ip address. The rationale behind this is that in this way you can easily setup the network address and mask for the VMs. To improve this, we can extend the parameters to define the ranged network, e.g. with the starting ip address parameter. Or alternatively, define a fixed net using single leases and/or continuous ranges of leases. Which one is better? BTW, if the MAC is not specified OpenNebula generates one for you using the prefix in oned.conf and the 32 bits of the IP address (as shown in you shell script). Cheers Ruben On Thu, Apr 28, 2011 at 7:49 PM, Lars Kellogg-Stedman l...@seas.harvard.edu wrote: we have setup up a ranged virtual network for our ONE-2.2 VMs but the gateway IP address keeps being assigned to VMs. Ahmad, I just asked a very similar question on the list. I agree that the existing documentation isn't particularly clear. The only response I received was a suggestion to use FIXED networks rather than RANGED networks, and then to programmatically generate the list of leases. So our network configuration looks like this: BRIDGE=br619 NAME=Cloud Private TYPE=FIXED Followed by the output of the following shell script: for (( ip=20; ip255; ip++ )); do printf LEASES = [IP=10.243.19.%d, MAC=5E:A5:%02X:%02X:%02X:%02X]\n $ip 10 243 19 $ip done This generates output that looks like this: LEASES = [IP=10.243.19.20, MAC=5E:A5:0A:F3:13:14] LEASES = [IP=10.243.19.21, MAC=5E:A5:0A:F3:13:15] [...] LEASES = [IP=10.243.19.254, MAC=5E:A5:0A:F3:13:FE] In this case, we're using the MAC prefix model, so we're able to pre-generate this list. This has solved the problem of OpenNebula handing out the router address to virtual machines, at the cost of a small amount of complexity. -- Lars Kellogg-Stedman l...@seas.harvard.edu Senior Technologist Harvard University SEAS Academic and Research Computing (ARC) ___ Users mailing list Users@lists.opennebula.org http://lists.opennebula.org/listinfo.cgi/users-opennebula.org -- Dr. Ruben Santiago Montero Associate Professor (Profesor Titular), Complutense University of Madrid URL: http://dsa-research.org/doku.php?id=people:ruben Weblog: http://blog.dsa-research.org/?author=7 ___ Users mailing list Users@lists.opennebula.org http://lists.opennebula.org/listinfo.cgi/users-opennebula.org