Perl 6 - the future is here, just unevenly distributed

IRC log for #fuel, 2015-08-04

| Channels | #fuel index | Today | | Search | Google Search | Plain-Text | summary

All times shown according to UTC.

Time Nick Message
00:16 tzn joined #fuel
00:52 alwaysatthenoc joined #fuel
01:31 wayneeseguin joined #fuel
01:36 xarses joined #fuel
01:52 ken316 Is it okay to map IP/Node/MAC for OpenStack nodes to be deployed prior to deployment?
01:54 skylerberg joined #fuel
02:04 ub joined #fuel
02:25 pmcg mwhahaha: it's most definitely this https://bugs.launchpad.net/fuel/+bug/1463035 (i know i said we were using classic, but i was incorrect)
02:41 monester joined #fuel
02:52 hakimo joined #fuel
03:13 eliqiao joined #fuel
03:13 eliqiao left #fuel
03:17 tzn joined #fuel
03:29 eliqiao joined #fuel
03:37 eliqiao1 joined #fuel
03:50 jhova joined #fuel
03:56 RedShift_ joined #fuel
04:21 sc68cal joined #fuel
04:21 sc68cal joined #fuel
04:30 neophy joined #fuel
04:56 monester joined #fuel
04:59 ximepa joined #fuel
05:29 sc68cal joined #fuel
05:35 Billias_ joined #fuel
05:37 bogdando joined #fuel
05:43 tzn joined #fuel
06:21 mkwiek joined #fuel
06:32 ximepa left #fuel
06:34 tzn joined #fuel
06:35 tzn joined #fuel
06:42 youellet joined #fuel
06:45 RedShift_ joined #fuel
07:08 kpimenova joined #fuel
07:39 HeOS joined #fuel
07:41 jesuspv joined #fuel
07:42 youellet_ joined #fuel
07:49 monester joined #fuel
07:49 ub joined #fuel
08:01 xenolog13 gcossu, You can try cherry-pick fix (commit 47425f336b4cdff3e452de067e8684cb1a64f5ab) to the 6.1 ns_IPaddr2 ocf script before start porting to 6.0. I guess this can fix you issue with ARP inconsistency.
08:03 RedShift_ joined #fuel
08:22 karume joined #fuel
08:28 JoeStack joined #fuel
08:37 karume joined #fuel
08:37 bogdando joined #fuel
08:37 wayneeseguin joined #fuel
08:37 Work^Seony joined #fuel
08:37 kaliya joined #fuel
08:37 mattymo joined #fuel
08:37 martineg_ joined #fuel
08:38 bearish joined #fuel
08:39 karume_ joined #fuel
08:44 e0ne joined #fuel
08:46 RedShift_ joined #fuel
08:52 kutija joined #fuel
09:09 RedShift_ joined #fuel
09:28 RedShift_ joined #fuel
09:59 karume joined #fuel
10:13 devvesa joined #fuel
10:14 JoeStack Is there an option to define a proxy the fuel will use? I tried to define it in /etc/environment, but the bootstrap_admin_node.sh could not resolve http://mirror.fuel-infra.org/mos/centos-6/mos6.1/updates
10:15 JoeStack From operating system level it works fine (yum update)
10:23 kutija joined #fuel
10:36 neophy joined #fuel
11:15 karume joined #fuel
11:22 kutija joined #fuel
11:52 tzn joined #fuel
11:59 jaypipes joined #fuel
12:33 karume joined #fuel
12:42 _tzn joined #fuel
12:48 neophy joined #fuel
12:50 monester joined #fuel
12:57 kutija joined #fuel
12:58 ken316 Is it okay to map IP/Node/MAC for OpenStack nodes to be deployed prior to deployment?
13:00 jaypipes joined #fuel
13:02 jdandrea joined #fuel
13:16 nurla joined #fuel
13:24 nihilifer joined #fuel
13:34 RedShift_ joined #fuel
13:43 thumpba joined #fuel
13:46 RedShift_ joined #fuel
14:03 ub joined #fuel
14:06 julien_ZTE joined #fuel
14:12 mudblur joined #fuel
14:12 nurla joined #fuel
14:15 karume joined #fuel
14:21 deckardkain hi guys. Was wondering if anybody used fuel to provide tenant instances direct vlan access via the public network. So instead of instances automatically getting a  private IP they would get an IP on the public network.
14:21 deckardkain i know there are articles on how to do this by manipulating the availbale bridge networks. But i was hoping to utilize fuel in order to scale out in the future.
14:25 mwhahaha you probably could get it to work
14:25 mwhahaha who would provide the public ip space?
14:25 mwhahaha the issue would be security concerns around allowing direct public ip access
14:25 mwhahaha is there a reason you wouldn't just do it via floating ips?
14:29 deckardkain mwhahaha: understood about the security issues. The reason why we want to do it, is a POC that openstack can integrate with our current environment. By using floating IPs we will be introducing double NAT in our environment. The provider of the public IP space will be provided by external means on an external router.
14:33 mwhahaha so there's no reason you couldn't just configure fuel to use your public ip space as the private network
14:33 mwhahaha because you just provide an ip block and gateway
14:33 mwhahaha but you could only scale your cloud as big as your public ip space that way
14:34 mwhahaha i don't think you'd be doing a double nat if you are using neutron vlans
14:34 mwhahaha maybe my understanding is wrong, but i don't think you end up with a double nat if you do it right
14:36 tzn joined #fuel
14:42 deckardkain mwhahaha: well the double nat happens on the "real" public ip space which is handled by an F5 which does a NAT to internal addresses. Those internal addresses would be in affect another NAT to the opnstack instance. Which our network team does not want to introduce into our environment.
14:44 mwhahaha well if you f5 nat to a range of internal IPs that you can use as the 'private' ip space for your openstack cluster then you only have the single nat
14:44 mwhahaha the f5 nat seems silly, but that's completely up to your network folks on how that gets utilized
14:46 deckardkain mwhahaha: hmm. don't think i've tried setting the private IP space as a real routable L3 network. I'll need to try that.
14:46 mwhahaha it's just IP space :)
14:46 mwhahaha in theory it'll work just fine
14:47 warpc__ joined #fuel
14:48 deckardkain mwhahaha: so for neutron L2 configuration section in Fuel. I would essentially just match the corresponding VLAN with that IP space?
14:48 mwhahaha yea
14:48 deckardkain mwhahaha: wouldn't the openstack router interefere with the tagged packet?
14:48 xarses deckardkain: it sounds like you want provider networks which fuel wont configure for you
14:49 deckardkain @xarses: hmm. So through fuel customization we won't be able to configure provider networks? My knowledge on provider networks is pretty sparse.
14:49 claflico joined #fuel
14:49 deckardkain @xarses: do you recommend any good reading?
14:50 xarses deckardkain: through settings, no fuel will not configure provider networks. The manifests technically could be forced to do that, but I'm not sure what would need to be set for that
14:50 mwhahaha looking at the network settings now, yea it won't work because i think we provide the internal network gateway which would conflict with what you'd need to set
14:51 tzn joined #fuel
14:54 deckardkain @xarses: do you think a fuel plugin could configure and create provider networks?
14:55 xarses deckardkain: in 7.0 sure, in 6.1 it will be more difficult
14:56 angdraug joined #fuel
14:56 deckardkain @xarses: do you have a link to a blue print that may relate, if there is one that exists?
14:57 RedShift_ joined #fuel
14:57 deckardkain i'm reading the openstack docs about provider networks, the more and more i read about them fits exactly what i need.
15:00 xarses deckardkain: there appear to be no fuel blueprints for them
15:00 deckardkain @xarses: ok. Then i'll assume that in fuel 7.0 it'll be a bit easier to write a fuel plugin in order to configure provider networks. Is that a fair assessment?
15:01 deckardkain generally speaking of course
15:01 xarses I'd need to see exactly what has to be done, In general plugins are going to be better with each release
15:01 xarses and therfor easier
15:02 deckardkain @xarses: gotcha.
15:02 xarses if its all config based, it would be about the same
15:02 xarses say in neutron
15:03 xarses if you need to mess with the network scheme and change the bridges around you need your plugin and 7.0
15:03 deckardkain got it
15:03 deckardkain @xarses: thank you so much for your contributions to openstack and for your time just now.
15:03 deckardkain mwhahaha: thanks for your time as well.
15:03 mwhahaha no problem
15:03 xarses deckardkain: let us know if you need help with the plugin
15:04 deckardkain will do
15:09 blahRus joined #fuel
15:11 e0ne joined #fuel
15:23 RedShift_ joined #fuel
15:27 RedShif__ joined #fuel
15:28 karume joined #fuel
15:30 RedShift_ joined #fuel
15:39 neophy joined #fuel
15:48 RedShift_ joined #fuel
15:49 RedShift_ joined #fuel
15:54 tatyana joined #fuel
15:56 RedShift_ joined #fuel
16:01 preilly joined #fuel
16:03 RedShif__ joined #fuel
16:03 Guest97200 joined #fuel
16:04 nihilifer joined #fuel
16:32 RedShift_ joined #fuel
16:45 sergmelikyan joined #fuel
16:46 skylerberg joined #fuel
16:53 julien_ZTE joined #fuel
16:55 dilyin joined #fuel
17:04 karume joined #fuel
17:05 e0ne joined #fuel
17:09 bildz hello
17:10 bildz i had to rebuild my fuel server again and am running into the duplicate IP problem like before
17:10 bildz what's the bug url for the fix
17:26 neophy joined #fuel
17:32 e0ne joined #fuel
17:40 neophy joined #fuel
17:58 nurla joined #fuel
17:58 mwhahaha was it this one? https://review.openstack.org/#/c/190089/
17:58 seens_ joined #fuel
17:58 monester joined #fuel
18:03 mudblur joined #fuel
18:13 bildz yes thank you!
18:14 nurla joined #fuel
18:46 monester joined #fuel
18:53 xarses joined #fuel
18:54 julien_ZTE joined #fuel
18:54 ub2 joined #fuel
18:54 xarses joined #fuel
19:01 monester joined #fuel
19:03 thansen joined #fuel
19:04 JoeStack Hi ! Is "yum update -y && dockerctl destroy all && bootstrap_admin_node.sh" the default update task?
19:12 ken316 I have deployed an 8-node cluster and saved the settings using "fuel [settings,network,node] --download .  I'd like to re-create it after rebuilding a Fuel-master using fuel..*--upload .  The "settings", and "network" worked, but the nodes appear to require discovery first.
19:12 ken316 Can I map a specific MAC to a specific node ID?
19:30 angdraug joined #fuel
19:36 xdeller joined #fuel
19:38 deckardkain i'm trying to installed openstack on virtual machines hosted on esxi. I'm trying to integrate with vCenter. My verify completes successfully, but when it gets to installed openstack on my nodes. I get the following error:
19:38 deckardkain http://paste.openstack.org/show/406932/
19:40 monester joined #fuel
19:43 mwhahaha deckardkain: what happens if you run that on node-18 manually?
19:43 mwhahaha do you get any error messages?
19:44 mwhahaha JoeStack: https://docs.mirantis.com/fuel/fuel-master/operations.html#applying-patches
19:44 mwhahaha ken316: not sure  you can do that, there is a backup/restore utility for the fuel master which might have needed to be run. I don't think it's a trival task to resetup the environment after the fact
19:44 deckardkain mwhahaha: oh heh i was trying to run it on the fuel master node let me run it on the node
19:45 tzn joined #fuel
19:49 ken316 okay, thanks.  And are coding things in plugins the best way to go whenever customizing (such as LDAP, Tenant list, KeyPairs) even though the intent is not to publish them (e.g. plugins)?
19:54 karume joined #fuel
19:57 mwhahaha ken316: probably because that way they will be automatically applied to any new systems you deploy. other wise you have to manually make whatever changes anytime you add new nodes. if you're ok with that then you can do that too
20:28 kutija joined #fuel
20:30 deckardkain mwhahaha: just following up.. Turns out from node-18 i didn't have DNS resolving correctly. When i added my vcenter server, i added the FQDN. To test the theory quickly i added a host entry in /etc/hosts on node-18 and redeployed and worked again.
20:31 mwhahaha cool, good to know you were able to continue
20:37 sergmelikyan joined #fuel
20:41 deckardkain mwhahaha: yup. pretty exciting
20:42 deckardkain mwhahaha: thanks again for your time
20:43 JoeStack <mwhahaha>  this documet was exactly the reason why i asked about the bootstrap_admin.sh command which is mentioned at first login to the fuel node.
20:45 tzn joined #fuel
20:45 HeOS joined #fuel
20:45 ken316 mwhahaha:  Thanks for you help!
20:51 sergmelikyan joined #fuel
20:53 mwhahaha JoeStack: the document should be to date, so i believe that is the correct set of steps
20:53 mwhahaha er be up to date
20:55 julien_ZTE joined #fuel
21:08 kpimenova joined #fuel
21:29 JoeStack My deployment was successful and I should be able to login to the dashbord at http://192.168.1.101 but nothing happens. In the L2 domain I see an arp for 192.168.1.101 include MAC, but no port seems to be open. Any idea?
21:34 JoeStack 192.168.1.0/24 is my public network. public is bind to eth2 on all nodes as native vlan.
21:36 JoeStack eth0 is admin(pxe) and eth1 for tagged vlans
21:38 thumpba joined #fuel
21:38 bildz How long does it usually take to deploy the base OS on the nodes?
21:39 JoeStack <bildz> up to 20-30 minutes in my deployment
21:40 bildz cool
21:40 bildz does the installing centos/ubuntu progress bar move or just stay at the start?
21:41 JoeStack centos moves faster but freezes at the end... ubuntu freezes at the beginning :-)
21:41 bildz how do you know if it's installing?
21:41 JoeStack click on log symbol from the specific node
21:42 JoeStack may choose debug as loglevel
21:43 karume joined #fuel
21:49 thumpba joined #fuel
21:52 tzn joined #fuel
21:56 thumpba_ joined #fuel
22:05 thumpba joined #fuel
22:09 thumpba joined #fuel
22:10 thumpba joined #fuel
22:26 julien_ZTE joined #fuel
22:55 tzn joined #fuel
23:32 karume joined #fuel
23:56 eliqiao joined #fuel
23:59 tzn joined #fuel

| Channels | #fuel index | Today | | Search | Google Search | Plain-Text | summary