Re: [Openstack] Per host floating ip
I have tried gluster for live migration but could net get it working. Please post back here if get that working. Instances ran fine with /var/lib/nova/instances on gluster, but live migration would change vm state to migrating and then back to running but still have it running where it was without any errors in the logs. M On Saturday, 7 July 2012, Alessandro Tagliapietra tagliapietra.alessan...@gmail.com wrote: I'm currently building up, i did a test setup with 3 machines and it worked. Now i'm doing it again with secure passwords, firewall etc. Btw, i just bought 3 EX 4 servers with a private lan between them using the switch + additional nic option they provide. I was thinking that another solution would be to use failover ip and when i assign an ip to an instance a custom hook calls the hetzner api to change the routing to point the ip to the server which is running the vm. Next step would be to use a shared fs like gluster for vm HA. I was thinking to make a blog post as soon as i get everything working to share how i did the setup. Regards Il giorno 06/lug/2012, alle ore 23:45, Michael Grosser ha scritto: Sounds like a solution, but a manual one. A more automatic way would be great I reckon. On Fri, Jul 6, 2012 at 7:00 PM, Marnus van Niekerk m...@mjvn.net wrote: I have limited knowledge of OpenStack, but I host many servers with Hetzner and know their infrastructure well. Since nova allows you to assign specific floating IPs to specific instances you could do the following. Create a floating range include all of the IPs Hetzner has given you. Then just associate the IP from the underlying host with a VM running on it that host using euca-associate-address. On 06/07/2012 18:17, Michael Grosser wrote: Hey, I would love to hear more on your setup at hetzner, I was looking to setup openstack there myself, but I didn't see a solution for the per server ip issue. How many machines do you use? How did you set it up? Do you use only nova or swift too? I would love to get more insight. Thanks for taking the time Cheers Michael On Fri, Jul 6, 2012 at 12:24 PM, Alessandro Tagliapietra tagliapietra.alessan...@gmail.com wrote: Hello guys, i've some servers at hetzner, i've actually installed essex on those and they're working fine, the step i'm missing is that hetzner gives you additional ip fixed to one host or failover ip that i can move to one server to another with the api. What i want is to assign floating ip to vm, depending on which host the vm is running or use a create hook to move the failover ip to the host which is running the vm. Is it possible in some way? Best Regards ___ Mailing list: https://launchpad.net/~openstack Post to : openstack@lists.launchpad.net Unsubscribe : https://launchpad.net/~openstack More help : https://help.launchpad.net/ListHelp ___ Mailing list: https://launchpad.net/~openstack Post to : openstack@lists.launchpad.net Unsubscribe : https://launchpad.net/~openstack More help : https://help.launchpad.net/ListHelp ___ Mailing list: https://launchpad.net/~openstack Post to : openstack@lists.launchpad.net Unsubscribe : https://launchpad.net/~openstack More help : https://help.launchpad.net/ListHelp
Re: [Openstack] PLEASE HELP multi_host not working
On 04/07/2012 15:53, Mark Russell wrote: There is likely a better way to fix this than what I did (drop and recreate nova db) but it was the simplest fix given that this was a brand new installation. There is almost certainly a better way to fix it by manually hacking nova db. Thank you Mark! Creating a new nova database called novaMulti and creating the pvt and floating networks from scratch has taken me leaps forward. Still struggling a bit with some of the routing, but at least I can now launce VMs on all of the compute nodes. Tx M ___ Mailing list: https://launchpad.net/~openstack Post to : openstack@lists.launchpad.net Unsubscribe : https://launchpad.net/~openstack More help : https://help.launchpad.net/ListHelp
Re: [Openstack] Per host floating ip
I have limited knowledge of OpenStack, but I host many servers with Hetzner and know their infrastructure well. Since nova allows you to assign specific floating IPs to specific instances you could do the following. Create a floating range include all of the IPs Hetzner has given you. Then just associate the IP from the underlying host with a VM running on it that host using euca-associate-address. On 06/07/2012 18:17, Michael Grosser wrote: Hey, I would love to hear more on your setup at hetzner, I was looking to setup openstack there myself, but I didn't see a solution for the per server ip issue. How many machines do you use? How did you set it up? Do you use only nova or swift too? I would love to get more insight. Thanks for taking the time Cheers Michael On Fri, Jul 6, 2012 at 12:24 PM, Alessandro Tagliapietra tagliapietra.alessan...@gmail.com wrote: Hello guys, i've some servers at hetzner, i've actually installed essex on those and they're working fine, the step i'm missing is that hetzner gives you additional ip fixed to one host or failover ip that i can move to one server to another with the api. What i want is to assign floating ip to vm, depending on which host the vm is running or use a create hook to move the failover ip to the host which is running the vm. Is it possible in some way? Best Regards ___ Mailing list: https://launchpad.net/~openstack Post to : openstack@lists.launchpad.net Unsubscribe : https://launchpad.net/~openstack More help : https://help.launchpad.net/ListHelp ___ Mailing list: https://launchpad.net/~openstack Post to : openstack@lists.launchpad.net Unsubscribe : https://launchpad.net/~openstack More help : https://help.launchpad.net/ListHelp ___ Mailing list: https://launchpad.net/~openstack Post to : openstack@lists.launchpad.net Unsubscribe : https://launchpad.net/~openstack More help : https://help.launchpad.net/ListHelp
Re: [Openstack] PLEASE HELP multi_host not working
On 04/07/2012 21:43, Vishvananda Ishaya wrote: It looks like the ip for your hosts is in the 10./8 range which is probably messing up routing and snatting. You will need to use a smaller range when you create your vm network, say 10.75.0.0/16 http://10.75.0.0/16 and make sure you set fixed_range to the same value in nova.conf I already have it sub-netted to 10.10.11.128/26 and have --fixed_range=10.10.11.128/26 in nova.conf. For metadata, you may have to set metadata_host (it sets this automatically, but if you have multiple interfaces it could be getting the wrong value) to the ip of the compute node on each compute host (and run nova-api-metadata or nova-api on each compute host as well. On each host I see an iptables rule like this natting the 169.254.169.254 address to the br100 address of the host like this: -A nova-network-PREROUTING -d 169.254.169.254/32 -p tcp -m tcp --dport 80 -j DNAT --to-destination 10.10.20.114:8775 But wget http://169.254.169.254/; on the controller node gives me the horizon index page! On the other compute nodes it gives connection refused. Should metadata_host be set to the public (eth0) ip of the host instead? Tx for the advice so far. M ___ Mailing list: https://launchpad.net/~openstack Post to : openstack@lists.launchpad.net Unsubscribe : https://launchpad.net/~openstack More help : https://help.launchpad.net/ListHelp
Re: [Openstack] multi_host not working
On 02/07/2012 16:33, Razique Mahroua wrote: I've put a small section here http://docs.openstack.org/diablo/openstack-compute/admin/content/multi-host.html Using this I have made progress, except I had to use nova-manage network delete 10.10.11.128/26 to delete the network and then added it back with the --multi_host=T option. I can now see the bridge created and assigned an address on each compute node, but all of the VMs get stuck after the bootloader - they never boot any further. What else could be wrong? Should nova-api be running on each compute node? Tx Marnus ___ Mailing list: https://launchpad.net/~openstack Post to : openstack@lists.launchpad.net Unsubscribe : https://launchpad.net/~openstack More help : https://help.launchpad.net/ListHelp
Re: [Openstack] multi_host not working
On 03/07/2012 09:53, Marnus van Niekerk wrote: I can now see the bridge created and assigned an address on each compute node, but all of the VMs get stuck after the bootloader - they never boot any further. Sorry, they do actually boot after a while but without any networking.. cloud-init-nonet waiting 120 seconds for a network device. cloud-init-nonet gave up waiting for a network device. ci-info: lo: 1 127.0.0.1 255.0.0.0 . ci-info: eth0 : 1 . . fa:16:3e:39:5f:02 route_info failed Waiting for network configuration... Waiting up to 60 more seconds for network configuration... Booting system without full network configuration... ___ Mailing list: https://launchpad.net/~openstack Post to : openstack@lists.launchpad.net Unsubscribe : https://launchpad.net/~openstack More help : https://help.launchpad.net/ListHelp
Re: [Openstack] [OpenStack][Nova] Live Migration + NFSv4 - Permission issues
Have you tried setting the ownership of /var/lib/nova/instances to the nova user? sudo chown -R nova:nova /var/lib/nova/instances M On 03/07/2012 15:48, Leander Bessa Beernaert wrote: Hello all, I've been trying to get the live migration to work according to the guide http://docs.openstack.org/trunk/openstack-compute/admin/content/configuring-live-migrations.html . So far i've setup 2 compute nodes and 1 controller node. They all share the /var/lib/nova/instances dir. I've already verified that the nova user id is the same across all the servers. Currently i'm running into this error when i launch an instance: http://paste.openstack.org/show/19221/ It's certainly a permission issue, so i tried adding the group nova to the user libvirt-qemu. However, it still doesn't work. To which user must i give the nova group permission in order to be able to write in that directory? Regards, Leander ___ Mailing list: https://launchpad.net/~openstack Post to : openstack@lists.launchpad.net Unsubscribe : https://launchpad.net/~openstack More help : https://help.launchpad.net/ListHelp ___ Mailing list: https://launchpad.net/~openstack Post to : openstack@lists.launchpad.net Unsubscribe : https://launchpad.net/~openstack More help : https://help.launchpad.net/ListHelp
Re: [Openstack] PLEASE HELP multi_host not working
I installed a minimal cloud image (cirros) and found that it had no network configures after booting. Even running udhcpc manually does not work - it never gets an answer from dhcp. I checked and dnsmasq is running on all of the the compute nodes: /usr/sbin/dnsmasq --strict-order --bind-interfaces --conf-file= --domain=novalocal --pid-file=/var/lib/nova/networks/nova-br100.pid --listen-address=10.10.11.135 --except-interface=lo --dhcp-range=10.10.11.130,static,120s --dhcp-lease-max=64 --dhcp-hostsfile=/var/lib/nova/networks/nova-br100.conf --dhcp-script=/usr/bin/nova-dhcpbridge --leasefile-ro Why would the DHCP not be working? I am running out of ideas here. Have tried everything I can read from searching google and still multi_host does not come close to working. I am now even considering abandoning OpenStack and going back to Eucalyptus. I have even tried changing to FlatNetwork with flat_injected=True and then IPs are set in the VMs but the metadata service still fails (no route to host). I have checked and each host has the 169.254.269.254 addresses natted to itself. I have also deleted and recreated the network with the --multi_host=T flag. I have this on the controller: #--routing_source_ip=10.10.20.114 --ec2_dmz_host=10.10.20.114 --multi_host=True --enabled_apis=ec2,osapi_compute,osapi_volume,metadata --dmz_cidr=10.10.20.114/32 And this on the compute nodes: #--routing_source_ip=10.10.20.114 --ec2_dmz_host=10.10.20.114 --multi_host=True --enabled_apis=metadata --dmz_cidr=10.10.20.114/32 On 03/07/2012 10:03, Marnus van Niekerk wrote: On 03/07/2012 09:53, Marnus van Niekerk wrote: I can now see the bridge created and assigned an address on each compute node, but all of the VMs get stuck after the bootloader - they never boot any further. Sorry, they do actually boot after a while but without any networking.. cloud-init-nonet waiting 120 seconds for a network device. cloud-init-nonet gave up waiting for a network device. ci-info: lo: 1 127.0.0.1 255.0.0.0 . ci-info: eth0 : 1 . . fa:16:3e:39:5f:02 route_info failed Waiting for network configuration... Waiting up to 60 more seconds for network configuration... Booting system without full network configuration... ___ Mailing list: https://launchpad.net/~openstack Post to : openstack@lists.launchpad.net Unsubscribe : https://launchpad.net/~openstack More help : https://help.launchpad.net/ListHelp
[Openstack] multi_host not working
Hi. I am trying to use multi_host to eliminate the controller hosts as a single point of failure. I followed the steps at http://docs.openstack.org/essex/openstack-compute/admin/content/existing-ha-networking-options.html and added thse options to the end of nova.conf. Now the guests have no connectivity to the outside world at all. (Running on ubuntu 12.04 using packages.) Controller: --multi_host=True --enabled_apis=ec2,osapi_compute,osapi_volume,metadata Compute nodes: --multi_host=True --enabled_apis=metadata I also tried changing the routing_source_ip option on each compute node to it's own ip address but it makes no difference. --routing_source_ip=10.10.20.11X What am I missing? Tx Marnus van Niekerk ___ Mailing list: https://launchpad.net/~openstack Post to : openstack@lists.launchpad.net Unsubscribe : https://launchpad.net/~openstack More help : https://help.launchpad.net/ListHelp
Re: [Openstack] multi_host not working
I have managed to get this working by changing the default gateway on the guest to the compute node it is running on. ubuntu@monitor:~$ sudo route del default gw 10.10.11.129 ubuntu@monitor:~$ sudo route add default gw 10.10.11.112 But the default gateway is assigned by DHCP - so how can I change the default gateway that nova-network assigns on each compute node? Tx M On 02/07/2012 13:50, Marnus van Niekerk wrote: Hi. I am trying to use multi_host to eliminate the controller hosts as a single point of failure. I followed the steps at http://docs.openstack.org/essex/openstack-compute/admin/content/existing-ha-networking-options.html and added thse options to the end of nova.conf. Now the guests have no connectivity to the outside world at all. (Running on ubuntu 12.04 using packages.) Controller: --multi_host=True --enabled_apis=ec2,osapi_compute,osapi_volume,metadata Compute nodes: --multi_host=True --enabled_apis=metadata I also tried changing the routing_source_ip option on each compute node to it's own ip address but it makes no difference. --routing_source_ip=10.10.20.11X What am I missing? Tx Marnus van Niekerk ___ Mailing list: https://launchpad.net/~openstack Post to : openstack@lists.launchpad.net Unsubscribe : https://launchpad.net/~openstack More help : https://help.launchpad.net/ListHelp
Re: [Openstack] multi_host not working
On 02/07/2012 16:33, Razique Mahroua wrote: I've put a small section here http://docs.openstack.org/diablo/openstack-compute/admin/content/multi-host.html If I read this right then it is saying that nova-api should not run on every compute node, only nova-network and nova-compute. Is that right? Tx M ___ Mailing list: https://launchpad.net/~openstack Post to : openstack@lists.launchpad.net Unsubscribe : https://launchpad.net/~openstack More help : https://help.launchpad.net/ListHelp
Re: [Openstack] multi_host not working
On 02/07/2012 16:14, Nathanael Burton wrote: This is actually what multi_host should be doing when enabled. What node is that original gateway address from? Is that a different compute node? Yes, its is the br100 ip of the controller node which is also a compute node. M ___ Mailing list: https://launchpad.net/~openstack Post to : openstack@lists.launchpad.net Unsubscribe : https://launchpad.net/~openstack More help : https://help.launchpad.net/ListHelp
Re: [Openstack] multi_host not working
They are identical except for these options: --vncserver_proxyclient_address=10.10.20.11X --vncserver_listen=10.10.20.11X --routing_source_ip=10.10.20.11X The ip that gets issues as gateway is the one in this option --flat_network_dhcp_start=10.10.11.129 The network range for the flat network and the underlying network on the 2nd NIC is the same 10.10.11.0/24. Not sure if this should be this way or if it is better for the flat network with a different range. Each of the nodes have two NICs: eth0 - 10.10.20.11X - public eth3 - 10.10.11.11X - private - bridged to br100 and assigned 10.10.11.129 on the controller node They also have eth1 and eth2 bonded into bond0 with IP 10.10.12.11X, but that is not used by OpenStack at all. Tx M On 02/07/2012 17:02, Nathanael Burton wrote: Are the nova.conf files identical across all the nodes? On Jul 2, 2012 10:47 AM, Marnus van Niekerk m...@mjvn.net mailto:m...@mjvn.net wrote: On 02/07/2012 16:14, Nathanael Burton wrote: This is actually what multi_host should be doing when enabled. What node is that original gateway address from? Is that a different compute node? Yes, its is the br100 ip of the controller node which is also a compute node. M ___ Mailing list: https://launchpad.net/~openstack Post to : openstack@lists.launchpad.net Unsubscribe : https://launchpad.net/~openstack More help : https://help.launchpad.net/ListHelp
[Openstack] Access a 3rd network
Hi, I'm new here so if this is a stupid question or the wrong place to ask, please just point me in the right direction. I have nova-compute running on 4 hosts with nova-network running on the controller host using FlatNetwork. Each of the hosts have three network interfaces: eth0 - 10.10.20.0/24 - public eth3 - 10.10.11.0/24 - private/bridge bond0 - 10.10.12.0/24 - eth1 and eth2 bonded together - not used by nova at the moment The VMs have access to the 10.10.20.0/24 network and the world via floating ip's, but I also need them to have access to the 10.10.12.0/24 network. How can this be done? Is there a way to create a 2nd public network bridged/natted to the bond0 interface? I tried doing this manually by adding an ip to the bond0 interface and the iptables rules, but that only worked from the 12 network to the VMs not from the VMs to the 12 network. Thank you Marnus van Niekerk ___ Mailing list: https://launchpad.net/~openstack Post to : openstack@lists.launchpad.net Unsubscribe : https://launchpad.net/~openstack More help : https://help.launchpad.net/ListHelp
[Openstack] Access a 3rd network
Hi, I'm new here so if this is a stupid question or the wrong place to ask, please just point me in the right direction. I have nova-compute running on 4 hosts with nova-network running on the controller host using FlatNetwork. Each of the hosts have three network interfaces: eth0 - 10.10.20.0/24 - public eth3 - 10.10.11.0/24 - private/bridge bond0 - 10.10.12.0/24 - eth1 and eth2 bonded together - not used by nova at the moment The VMs have access to the 10.10.20.0/24 network and the world via floating ip's, but I also need them to have access to the 10.10.12.0/24 network. How can this be done? Is there a way to create a 2nd public network bridged/natted to the bond0 interface? I tried doing this manually by adding an ip to the bond0 interface and the iptables rules, but that only worked from the 12 network to the VMs not from the VMs to the 12 network. Thank you Marnus van Niekerk ___ Mailing list: https://launchpad.net/~openstack Post to : openstack@lists.launchpad.net Unsubscribe : https://launchpad.net/~openstack More help : https://help.launchpad.net/ListHelp