| *** markvoelker has joined #openstack-ansible | 00:13 | |
| *** markvoelker has quit IRC | 00:18 | |
| *** JRobinson__ has quit IRC | 00:19 | |
| *** dkalleg has quit IRC | 00:45 | |
| *** galstrom_zzz is now known as galstrom | 00:47 | |
| *** dkalleg has joined #openstack-ansible | 01:45 | |
| *** dkalleg has quit IRC | 01:50 | |
| *** galstrom is now known as galstrom_zzz | 01:57 | |
| *** markvoelker has joined #openstack-ansible | 02:02 | |
| *** radek_ has joined #openstack-ansible | 02:05 | |
| *** markvoelker has quit IRC | 02:07 | |
| *** radek_ has quit IRC | 02:17 | |
| *** daneyon_ has quit IRC | 02:29 | |
| *** javeriak has quit IRC | 02:36 | |
| *** stevemar has quit IRC | 02:40 | |
| *** galstrom_zzz is now known as galstrom | 02:45 | |
| *** JRobinson__ has joined #openstack-ansible | 03:23 | |
| *** markvoelker has joined #openstack-ansible | 03:50 | |
| *** markvoelker has quit IRC | 03:55 | |
| *** galstrom is now known as galstrom_zzz | 04:32 | |
| *** markvoelker has joined #openstack-ansible | 05:39 | |
| *** markvoelker has quit IRC | 05:44 | |
| evrardjp | morning everyone | 06:56 |
|---|---|---|
| svg | Hi all | 07:15 |
| *** javeriak has joined #openstack-ansible | 07:18 | |
| *** markvoelker has joined #openstack-ansible | 07:28 | |
| *** JRobinson__ has quit IRC | 07:28 | |
| *** markvoelker has quit IRC | 07:33 | |
| evrardjp | hello svg | 07:34 |
| * svg still does death dances around network interfaces | 07:36 | |
| odyssey4me | lol, good morning | 07:37 |
| *** javeriak has quit IRC | 07:37 | |
| evrardjp | svg, for that part, I don't know if it could be related, but did you try load testing with something like iperf on your hardware? | 07:39 |
| evrardjp | just to see if everything is fine | 07:39 |
| evrardjp | even on containers | 07:39 |
| evrardjp | and in both ways | 07:39 |
| svg | actuall couple of colleagues are testing the hell out of it | 07:43 |
| svg | so far no cigar | 07:43 |
| svg | The thing we are wondering now, is something show in the exampple interfaces config in osad | 08:05 |
| svg | it show bond1.1998 being member of br-vxlan, and bond1 being member of br-vlan | 08:06 |
| svg | but there are lots of references that explain this cannot work: combining a vlan tagged interface in a bridge and that same interface untagged in another bridge | 08:08 |
| svg | see https://wiki.debian.org/NetworkConfiguration#Caveats_when_using_bridging_and_vlan | 08:09 |
| svg | we have our intrfaces setup as per default in that example, except we use only 1 bond0 | 08:09 |
| odyssey4me | svg it's possible that the example file has some mistakes | 08:14 |
| svg | perhaps, but then I'm not sure how we would be supposed to configure br-vlan | 08:15 |
| svg | except for using a physicaly separate interface/bond | 08:16 |
| svg | (if this is true) | 08:16 |
| evrardjp | is this what the example shows? | 08:19 |
| evrardjp | I thought the bond was only for physical layer, and they used vlans on top to configure br-vlan and br-vxlan | 08:19 |
| svg | no, as per ^^ bond1 does bioth | 08:19 |
| evrardjp | Could you resend the link with the (possibly faulty) documentation? | 08:20 |
| evrardjp | http://docs.rackspace.com/rpc/api/v10/bk-rpc-installation/content/sec_overview_neutron.html | 08:21 |
| evrardjp | because when you look at that | 08:23 |
| evrardjp | http://docs.rackspace.com/rpc/api/v10/bk-rpc-installation/content/sec-hosts-target-network-refarch.html | 08:23 |
| evrardjp | it seems that it's not wrong | 08:23 |
| svg | https://github.com/stackforge/os-ansible-deployment/blob/kilo/etc/network/interfaces.d/openstack_interface.cfg.example | 08:23 |
| evrardjp | the vlan are defined on top of the bond | 08:23 |
| evrardjp | then bridged | 08:24 |
| svg | that is consistent with those docs yes | 08:24 |
| svg | I'm not saying iut is wrong, only that it seems wrong If that Debian doc is right | 08:24 |
| svg | https://wiki.debian.org/NetworkConfiguration#Caveats_when_using_bridging_and_vlan | 08:25 |
| svg | and it sounds like this might be the cause of our troubles | 08:25 |
| evrardjp | first, I'd like to say that sometimes ubuntu and debian are inconsistent in terms of documentation vs the actual implementation | 08:27 |
| evrardjp | did you try without the bonding? just to make sure it works? | 08:28 |
| evrardjp | I mean just the bridging the vifs with the NICs | 08:30 |
| evrardjp | it's maybe only bonding that's causing you the pain | 08:33 |
| *** sdake has quit IRC | 08:34 | |
| svg | we did try to keeponly one interface in the bond | 08:36 |
| svg | this rackspace document also cocnfirms this http://www.rackspace.com/blog/vms-vlans-and-bridges-oh-my-part-2/ | 08:39 |
| evrardjp | I meant not configuring the bond at all | 08:49 |
| evrardjp | the link you're giving is from 2012... I guess they were also changes in the ubuntu 14.04 release | 08:52 |
| *** markvoelker has joined #openstack-ansible | 09:17 | |
| *** markvoelker has quit IRC | 09:22 | |
| openstackgerrit | Merged stackforge/os-ansible-deployment: Allow the Horizon SESSION_TIMEOUT to be configured https://review.openstack.org/188140 | 09:24 |
| *** markvoelker has joined #openstack-ansible | 10:17 | |
| *** markvoelker has quit IRC | 10:23 | |
| *** markvoelker has joined #openstack-ansible | 11:18 | |
| *** markvoelker has quit IRC | 11:22 | |
| openstackgerrit | Merged stackforge/os-ansible-deployment: Allow the Horizon SESSION_TIMEOUT to be configured https://review.openstack.org/188144 | 11:54 |
| cloudnull | Morning. | 11:54 |
| cloudnull | Svg the file https://github.com/stackforge/os-ansible-deployment/blob/master/etc/network/interfaces.d/openstack_interface.cfg.example is the interface file that has been made generic but is very very close to what we have running in production, last I was aware of. | 11:58 |
| evrardjp | cloudnull: morning | 11:59 |
| odyssey4me | morning cloudnull :) | 11:59 |
| vincent_vdk | this is what i've always seen : evrardjp> I thought the bond was only for physical layer, and they used vlans on top to configure br-vlan and br-vxlan | 11:59 |
| cloudnull | In this case bond1 is serving vxlan and vlan using a tagged and untagged interface. | 11:59 |
| cloudnull | Morning gents. | 11:59 |
| vincent_vdk | good afternoon :) | 12:00 |
| cloudnull | That too ;) | 12:00 |
| cloudnull | With bond1 the untagged bridge is given directly to neutron for provider networks which it uses with its set vlan tags. | 12:01 |
| evrardjp | cloudnull: very close is not equal | 12:01 |
| evrardjp | and apparently ubuntu is kinda difficult with its /etc/network/interfaces | 12:02 |
| cloudnull | Its about 98% there. | 12:02 |
| cloudnull | We setup an additional storage network for backups. | 12:03 |
| evrardjp | svg: the last time I had an issue with my bonding interface, I had to upgrade the firmware of the NICs | 12:03 |
| evrardjp | it fixed everything | 12:03 |
| evrardjp | maybe it could help | 12:03 |
| evrardjp | maybe not | 12:03 |
| cloudnull | And if the customer wants flat networks we do a little something like https://github.com/stackforge/os-ansible-deployment/blob/master/etc/network/interfaces.d/aio_interfaces.cfg | 12:03 |
| cloudnull | For the for the br-vlan | 12:04 |
| cloudnull | To float a beth pair off the bridge for the flat network to consume. | 12:04 |
| evrardjp | interesting | 12:04 |
| cloudnull | Evrardjp we had that issue to when trying to use melnox NICs. | 12:04 |
| cloudnull | And so.e Intel models. | 12:05 |
| cloudnull | *some | 12:05 |
| evrardjp | too bad I didn't have your feedback at that time :p | 12:05 |
| cloudnull | Next time for sure ;) | 12:05 |
| evrardjp | anyway I was just trying to help | 12:05 |
| cloudnull | That's awesome. | 12:06 |
| cloudnull | I just came work from my slumber logged in and thought it good to jiberjaber about network interfaces in Ubuntu first thing. :) | 12:07 |
| cloudnull | On an unrelated note. | 12:08 |
| cloudnull | How's everyone doing on this fine friday ? | 12:08 |
| svg | well according to that rackspace site I linked ^^ one should not have a bond0 and a bond0.X in bridges | 12:09 |
| cloudnull | Apsu wrote that blog post. When he awakes from his slumber maybe he can chime in on it. | 12:09 |
| svg | but that is the case in the .example interfaces file | 12:10 |
| cloudnull | Also I remember that to be true for 12.04 , it required some ebtables mungering to make it go. | 12:11 |
| cloudnull | But in 14.04 we've not had issue with it. | 12:11 |
| cloudnull | Evrardjp did klevinstine or sam-i-am get the doc issues sorted ? | 12:14 |
| cloudnull | I failed to check back on that yesterday. | 12:14 |
| evrardjp | cloudnull: didn't check that atm | 12:14 |
| evrardjp | I'll do the followup | 12:14 |
| evrardjp | I may not do that immediately though | 12:15 |
| cloudnull | No worries. Much appreciated. | 12:15 |
| svg | anyway, our testing seems to not indicate we have issues with specificaaly this | 12:16 |
| svg | a coworker noticed the fan getrs hit when stuff happens on the management bridge | 12:16 |
| svg | he now split out the bond in bond 0 and 1 with each 1 interface, and put the management bridge on a separate bond, and at first (now on one machine) that seems to fix stuff | 12:17 |
| svg | (.. ongoing..) | 12:17 |
| evrardjp | it's (I think) the best practice to split tenant traffic and management traffic | 12:18 |
| odyssey4me | svg in a former life, we used to run everything on a bonded pair of 10GB NIC's... we still had eth0 as a 1GB NIC which was just used for automated builds and management | 12:19 |
| evrardjp | rolls royce would be to split management and storage, but yes, interfaces | 12:19 |
| svg | we have two 10gb nics in one bond0, so yes, same thing | 12:20 |
| svg | evrardjp: you mean split physically on separate interfaces? | 12:20 |
| evrardjp | Bonding is always better because it brings resiliency. But if you can afford separate interfaces for management and tenants traffic, it's really better | 12:22 |
| evrardjp | -but | 12:23 |
| *** sc has quit IRC | 12:23 | |
| *** sc has joined #openstack-ansible | 12:26 | |
| *** markvoelker has joined #openstack-ansible | 12:34 | |
| *** markvoelker has quit IRC | 12:38 | |
| svg | yes, but why? | 12:41 |
| svg | nm, I see why obviosuly | 12:41 |
| *** KLevenstein has joined #openstack-ansible | 13:33 | |
| *** jaypipes is now known as leakypipes | 13:43 | |
| openstackgerrit | Merged stackforge/os-ansible-deployment: Allow the Horizon SESSION_TIMEOUT to be configured https://review.openstack.org/188143 | 13:47 |
| *** Mudpuppy has joined #openstack-ansible | 13:54 | |
| *** Mudpuppy has quit IRC | 13:55 | |
| *** Mudpuppy_ has joined #openstack-ansible | 13:55 | |
| *** Mudpuppy_ is now known as Mudpuppy | 13:58 | |
| *** markvoelker has joined #openstack-ansible | 14:05 | |
| *** markvoelker has quit IRC | 14:09 | |
| *** sigmavirus24_awa is now known as sigmavirus24 | 14:12 | |
| openstackgerrit | git-harry proposed stackforge/os-ansible-deployment: Enable udev for lvm in cinder-volume container https://review.openstack.org/188797 | 14:33 |
| Apsu | svg: Yeah, that blog post applied to the kernels at the time. About a year later or so, they fixed the issue. | 14:35 |
| Apsu | I think they fixed it in... 3.11 or so. | 14:35 |
| svg | thx, good to know | 14:35 |
| cloudnull | boom the man the myth the legend Apsu in da house ! | 14:35 |
| Apsu | I was pretty surprised when it happened, lol | 14:35 |
| Apsu | cloudnull: ohai2uthar! | 14:36 |
| cloudnull | hows it brother ? | 14:36 |
| svg | either way, not sure if the part where mixing bond0 and bond0.x is bad still applies | 14:36 |
| Apsu | Not too bad. Clawing my way out of my melatonin-induced zombieism | 14:36 |
| openstackgerrit | Merged stackforge/os-ansible-deployment: Moved all of group_vars into a variable file https://review.openstack.org/188074 | 14:36 |
| svg | our tsesting didnt show that | 14:36 |
| Apsu | svg: It doesn't. | 14:36 |
| svg | so far it seems we have a problem having the mgmt bridge on the same bond as other things, which is very weird | 14:37 |
| svg | Apsu: do you have some 'official' reference to that? | 14:37 |
| Apsu | What do you mean exactly? | 14:37 |
| Apsu | svg: I can try to find a patch reference | 14:38 |
| svg | so originally we have a single bond0, which is used for all the rbdiges | 14:38 |
| svg | and we experience packet loss | 14:38 |
| svg | which was seen at deploy time a bit, but becomes very obvious once we start e.g. launching 100 vm instances | 14:39 |
| Apsu | Ok. | 14:39 |
| svg | when removing all the container veth interfaces from the mgmt bridge, the packet loss disappears | 14:39 |
| Apsu | What kernel? What NICs? | 14:40 |
| Apsu | What bonding mode? | 14:40 |
| svg | 3.13.0-46-generic, bond mode 1 (active passive), and nics are Intel 10GB (AOC-XEH-iN2, http://supermicro.com/products/superblade/networking/AOC-XEH-iN2.cfm) | 14:42 |
| svg | the current testing-workaround was to take one nic out of the bond, create a second bond1 with it, and put that (vlantagged) into the mgmt bridge interface instead of its bond0 counterpart | 14:43 |
| svg | with that setup we have no packet loss whatsovere | 14:43 |
| *** stevemar has joined #openstack-ansible | 14:44 | |
| Apsu | Hrm. | 14:44 |
| svg | Yes. | 14:45 |
| Apsu | Did you check the bond stats and individual link stats to see if it was one particular NIC misbehaving? | 14:47 |
| Apsu | Y'know, cat /proc/net/bonding/bond0, and ip -s -s link show $nic1, ... $nic2 | 14:48 |
| svg | not sure (didnt troublkeshoot this myself) but I believe we had this when any of the nics was primary | 14:50 |
| andymccr | once cinder 1.2.2 client is released we need to update the openstack_clients.yml :) | 14:51 |
| svg | What we did see, is lots of dropped packets on the passive nic | 14:53 |
| Apsu | svg: That's particularly strange, since it shouldn't be passing any :P | 14:53 |
| svg | yes | 14:53 |
| Apsu | Assuming it's actually passive | 14:53 |
| Apsu | lol | 14:53 |
| svg | that's what /proc/net/bonding/bond0 says :) | 14:54 |
| Apsu | Pretty good sign then | 14:57 |
| Apsu | How are you detecting packet loss on the passive link? | 14:57 |
| svg | dropped packets, on the ifconfig output | 15:02 |
| Apsu | Can you show me the output of ip -s -s link show $link? | 15:08 |
| Apsu | Also, in a network arch meeting, so afkish | 15:08 |
| *** cbaesema has quit IRC | 15:17 | |
| evrardjp | svg: another tool you could use is mii_tool | 15:29 |
| evrardjp | to see the current mii status | 15:29 |
| evrardjp | anyway I'm off for today, see you next week! | 15:30 |
| *** markvoelker has joined #openstack-ansible | 15:30 | |
| *** markvoelker has quit IRC | 15:34 | |
| *** daneyon has joined #openstack-ansible | 15:36 | |
| *** daneyon has quit IRC | 15:38 | |
| cloudnull | later evrardjp, have a good weekend | 15:38 |
| *** daneyon has joined #openstack-ansible | 15:38 | |
| openstackgerrit | git-harry proposed stackforge/os-ansible-deployment: Enable udev for lvm in cinder-volume container https://review.openstack.org/188394 | 15:41 |
| svg | Apsu: http://sprunge.us/PjPX | 15:41 |
| *** faman has joined #openstack-ansible | 15:51 | |
| *** faman has quit IRC | 15:53 | |
| *** famanfortune has joined #openstack-ansible | 15:55 | |
| Apsu | svg: So, those drops on your p1p2 bond slave are all RX | 15:59 |
| Apsu | And it's basically the same # of total RX packets | 16:00 |
| Apsu | Which means something is sending traffic into it from the switch. | 16:00 |
| Apsu | Presumably because it's got the same MAC address... now I thought that mode 1 bonding didn't team the MACs up to be the same between active and passive slaves | 16:01 |
| Apsu | svg: Can you paste the output of cat /proc/net/bonding/bond0? :D | 16:01 |
| Apsu | Also, are these links hooked into the same switch...? | 16:02 |
| svg | http://paste.ubuntu.com/11590869/ | 16:04 |
| svg | these are blades, interfaces are connected to two separate internal chassis swicthes | 16:05 |
| svg | AFAIK those don;t support LACP which is why we do active passive | 16:05 |
| svg | but I don;t know all detaiuls there | 16:06 |
| svg | (atm on train, flacky connection) | 16:06 |
| *** dkalleg has joined #openstack-ansible | 16:08 | |
| Apsu | kk | 16:08 |
| Apsu | svg: Ok, so the bonding driver knows the links have different MACs, it's just displaying them as teamed because of the bond. Cool. | 16:09 |
| Apsu | So in theory the switches shouldn't be sending traffic on both links for active/backup mode... | 16:10 |
| Apsu | But that's where those drops are coming in | 16:10 |
| Apsu | When using mode 1 bonding on Linux, you generally don't tell the switches anything at all about it | 16:10 |
| *** dkalleg has quit IRC | 16:12 | |
| svg | so basically the lost packets would be caused by the switch sending packets to the wrong port, which would mean it somehow got the wrong ip address? | 16:15 |
| svg | I have been wondering where those Link Failure Counts come from also | 16:16 |
| Apsu | No it's not the wrong IP | 16:16 |
| Apsu | This is purely a layer 2 thing | 16:16 |
| svg | sorry I meant mac address | 16:16 |
| Apsu | Essentially it means the switch is sending frames to the MAC, because it saw that MAC on frames entering that switch port, for the backup link | 16:17 |
| Apsu | Maybe your blade switches aren't communicating their MAC tables with each other very well. | 16:17 |
| Apsu | Or at all | 16:17 |
| Apsu | So they're operating as independent switches, and incoming traffic to your bond MAC is being sent to both bond slaves always :P | 16:17 |
| svg | I didn't think they were independent; that would be #$*&@ | 16:18 |
| Apsu | The number of RX frames on both slaves is almost the same, with a bit more on the active link | 16:18 |
| Apsu | So it's nearly what I said, that all incoming traffic comes to both links | 16:19 |
| Apsu | Very odd. | 16:19 |
| svg | so the so called workaround of having the interfaces split out in separate bonds just happens to solve this as there is only one interface in the bonds | 16:20 |
| svg | which as nothing to do with manage,ment bridge | 16:20 |
| svg | management bridge caries the most traffic when we start to deploy and configure things, so that might just be more obvious there | 16:21 |
| Apsu | Yep | 16:21 |
| Apsu | You're probably getting actual communication issues -- which wouldn't happen if it were just mirroring all traffic to both links, and the passive slave was dropping it all -- because it's not mirroring | 16:22 |
| Apsu | It's probably alternating or something similar, and the passive drops everything | 16:22 |
| Apsu | So the active has to constantly request TCP retransmits and shrink window sizes, slowing traffic to a crawl | 16:23 |
| Apsu | Which fails when you have a ton of deployment traffic :D | 16:23 |
| Apsu | yay networking | 16:23 |
| *** javeriak has joined #openstack-ansible | 16:24 | |
| Apsu | I bet the blade switches are somewhat independent, and setup active-active | 16:24 |
| Apsu | Which is either actively balancing traffic by MAC or or the MAC tables are flapping thinking the same MAC is on both switches | 16:25 |
| Apsu | Are you certain they don't support MLAG LACP? | 16:25 |
| svg | thats what the network guy told me | 16:27 |
| Apsu | svg: You could always verify by tcpdump/wireshark/tshark and check for a bunch of duplicates/retransmits/window shrinking events | 16:27 |
| svg | I think I should have a look at it myself | 16:27 |
| Apsu | There are also other bonding modes that may be appropriate if it doesn't do MLAG | 16:27 |
| *** javeriak_ has joined #openstack-ansible | 16:29 | |
| *** javeriak has quit IRC | 16:29 | |
| svg | you mean better appropriate in this case than mode1? | 16:32 |
| *** javeriak has joined #openstack-ansible | 16:35 | |
| Apsu | Right | 16:37 |
| Apsu | Depending on how your switches are actually behaving, mode0 or mode6 may be better. | 16:37 |
| Apsu | mode6 (ALB) has some other considerations to take into account | 16:37 |
| *** sdake has joined #openstack-ansible | 16:39 | |
| *** javeriak_ has quit IRC | 16:40 | |
| *** sdake_ has joined #openstack-ansible | 16:44 | |
| *** sdake has quit IRC | 16:48 | |
| *** famanfortune has quit IRC | 17:03 | |
| *** javeriak_ has joined #openstack-ansible | 17:09 | |
| *** javeriak has quit IRC | 17:13 | |
| *** dkalleg has joined #openstack-ansible | 17:18 | |
| *** markvoelker has joined #openstack-ansible | 17:19 | |
| *** markvoelker has quit IRC | 17:24 | |
| *** famanfortune has joined #openstack-ansible | 17:25 | |
| *** galstrom_zzz is now known as galstrom | 17:33 | |
| *** javeriak_ has quit IRC | 17:35 | |
| *** galstrom is now known as galstrom_zzz | 17:38 | |
| *** galstrom_zzz is now known as galstrom | 17:41 | |
| *** galstrom is now known as galstrom_zzz | 17:44 | |
| *** sdake has joined #openstack-ansible | 17:50 | |
| *** sdake_ has quit IRC | 17:54 | |
| *** sdake_ has joined #openstack-ansible | 17:56 | |
| *** jwagner_away is now known as jwagner | 17:59 | |
| *** sdake has quit IRC | 17:59 | |
| *** javeriak has joined #openstack-ansible | 18:04 | |
| *** sdake_ is now known as sdake | 18:10 | |
| *** galstrom_zzz is now known as galstrom | 18:11 | |
| *** serverascode has quit IRC | 18:54 | |
| *** metral is now known as metral_zzz | 18:55 | |
| *** cbaesema has joined #openstack-ansible | 19:07 | |
| *** sdake has quit IRC | 19:08 | |
| *** markvoelker has joined #openstack-ansible | 19:08 | |
| *** metral_zzz is now known as metral | 19:09 | |
| *** markvoelker has quit IRC | 19:12 | |
| *** abitha has joined #openstack-ansible | 19:28 | |
| abitha | hi, I am using os_image module to upload an image to glance and having issues to figure out the right path for the file to exist. should it exist in some specific path in ansible project , or should it be copied manually to the target machine and ran from there? | 19:32 |
| *** leakypipes has quit IRC | 19:34 | |
| cloudnull | if the image was uploaded to glance and the task succeeded and your using the glance file backend you should be able to find the image at /var/lib/glance/images | 19:35 |
| cloudnull | or something similar | 19:35 |
| abitha | this is when using create_image, while specifying the image 'filename', its not able to access an image that is in the ansible project structure | 19:37 |
| abitha | but if I copy the image to the target machine and give the absolute path it works fine. is this how its supposed to be? | 19:38 |
| cloudnull | to be quite honest idk , this is channel is geared at the os-ansible-deployment project | 19:39 |
| abitha | oh ok. thanks | 19:39 |
| cloudnull | maybe give http://docs.ansible.com/glance_image_module.html a go | 19:40 |
| cloudnull | that module has the "file" attribute which should be able to follow your local directory structure and upload the miage. | 19:40 |
| cloudnull | *image | 19:41 |
| abitha | thanks @cloudnull | 19:41 |
| *** moravec has quit IRC | 19:46 | |
| *** moravec has joined #openstack-ansible | 19:47 | |
| *** abitha has left #openstack-ansible | 19:55 | |
| openstackgerrit | Merged stackforge/os-ansible-deployment: Enable udev for lvm in cinder-volume container https://review.openstack.org/188797 | 20:05 |
| *** javeriak has quit IRC | 20:18 | |
| cloudnull | stevelle: https://github.com/cloudnull/osad-basic-test-node/blob/master/osad-basic-test-aio-playbook.yml#L94 | 20:19 |
| *** metral is now known as metral_zzz | 20:22 | |
| cloudnull | stevelle: sigmavirus24: https://gist.github.com/cloudnull/7325348 | 20:23 |
| *** markvoelker has joined #openstack-ansible | 20:24 | |
| *** metral_zzz is now known as metral | 20:27 | |
| *** markvoelker has quit IRC | 20:28 | |
| *** annashen has joined #openstack-ansible | 20:34 | |
| *** famanfortune has left #openstack-ansible | 20:36 | |
| *** famanfortune has joined #openstack-ansible | 20:43 | |
| *** serverascode has joined #openstack-ansible | 20:50 | |
| openstackgerrit | Steve Lewis proposed stackforge/os-ansible-deployment: Configure DB addresses for each service https://review.openstack.org/188926 | 21:13 |
| openstackgerrit | Kevin Carter proposed stackforge/os-ansible-deployment: Prep for 10.1.8 tag https://review.openstack.org/188173 | 21:15 |
| *** toddnni has joined #openstack-ansible | 21:16 | |
| *** Mudpuppy_ has joined #openstack-ansible | 21:29 | |
| *** Mudpuppy has quit IRC | 21:32 | |
| *** Mudpuppy_ has quit IRC | 21:33 | |
| openstackgerrit | git-harry proposed stackforge/os-ansible-deployment: Enable udev for lvm in cinder-volume container https://review.openstack.org/188394 | 21:34 |
| *** openstack has joined #openstack-ansible | 21:43 | |
| openstackgerrit | Kevin Carter proposed stackforge/os-ansible-deployment: Updated all of master to the latest SHAs https://review.openstack.org/188937 | 21:51 |
| *** javeriak has joined #openstack-ansible | 21:55 | |
| *** KLevenstein has quit IRC | 22:06 | |
| *** stevemar has quit IRC | 22:12 | |
| *** markvoelker has joined #openstack-ansible | 22:12 | |
| *** markvoelker has quit IRC | 22:17 | |
| *** annashen has quit IRC | 22:19 | |
| openstackgerrit | Kevin Carter proposed stackforge/os-ansible-deployment: Updated kilo for new dev work https://review.openstack.org/188954 | 22:22 |
| * svg proposes clouds without networking | 22:28 | |
| stevelle | +1 | 22:28 |
| stevelle | major boon for security there | 22:29 |
| cloudnull | according to defcore you dont have to have networking to be an OpenStack cloud . | 22:29 |
| * svg proposes ethernet switches with one single, simple, standardized configuration system | 22:31 | |
| svg | those should have neat html5 interfaces | 22:32 |
| *** daneyon has quit IRC | 22:33 | |
| *** daneyon has joined #openstack-ansible | 22:33 | |
| *** daneyon has quit IRC | 22:33 | |
| cloudnull | lol | 22:34 |
| *** daneyon has joined #openstack-ansible | 22:34 | |
| *** openstackgerrit has quit IRC | 22:37 | |
| *** openstackgerrit has joined #openstack-ansible | 22:37 | |
| openstackgerrit | Steve Lewis proposed stackforge/os-ansible-deployment: Configure DB addresses for each service https://review.openstack.org/188926 | 22:46 |
| *** famanfortune has quit IRC | 22:46 | |
| openstackgerrit | Javeria Khan proposed stackforge/os-ansible-deployment: Adding PLUMgrid plugin option to neutron https://review.openstack.org/188964 | 22:54 |
| javeriak | guys, i went ahead and commited the changes ^, for the plumgrid BP, hope this gives you more clarity on what it was asking for | 22:55 |
| *** annashen has joined #openstack-ansible | 23:20 | |
| openstackgerrit | Merged stackforge/os-ansible-deployment: Update for 11.0.2 release https://review.openstack.org/188934 | 23:21 |
| openstackgerrit | Merged stackforge/os-ansible-deployment: Prep for 10.1.8 tag https://review.openstack.org/188173 | 23:21 |
| *** sdake has joined #openstack-ansible | 23:26 | |
| *** annashen has quit IRC | 23:26 | |
| openstackgerrit | Kevin Carter proposed stackforge/os-ansible-deployment: Updated kilo for new dev work https://review.openstack.org/188954 | 23:31 |
| *** daneyon has quit IRC | 23:34 | |
| *** dkalleg has quit IRC | 23:52 | |
| *** sdake_ has joined #openstack-ansible | 23:57 | |
| *** galstrom is now known as galstrom_zzz | 23:59 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!