| *** LinStatSDR has joined #openstack-ansible | 00:11 | |
| cloudnull | xdfil: I don't believe so. should be ok to randomize and go. | 00:15 |
|---|---|---|
| openstackgerrit | Merged openstack/openstack-ansible-os_cinder master: Add support for cinder v3 api https://review.openstack.org/446503 | 00:16 |
| *** LinStatSDR has left #openstack-ansible | 00:18 | |
| *** woodard has quit IRC | 00:21 | |
| *** woodard has joined #openstack-ansible | 00:21 | |
| openstackgerrit | Kevin Carter (cloudnull) proposed openstack/openstack-ansible-ops master: define the default kernel statically https://review.openstack.org/446790 | 00:24 |
| *** Jeffrey4l has quit IRC | 00:25 | |
| *** jrobinson has quit IRC | 00:27 | |
| *** weezS has quit IRC | 00:35 | |
| *** sukesh has quit IRC | 00:38 | |
| *** gouthamr has joined #openstack-ansible | 00:38 | |
| *** deadnull has quit IRC | 00:48 | |
| *** acormier has joined #openstack-ansible | 00:53 | |
| *** SerenaFeng has joined #openstack-ansible | 00:59 | |
| *** cuongnv has joined #openstack-ansible | 01:00 | |
| *** MasterOfBugs has quit IRC | 01:01 | |
| *** acormier has quit IRC | 01:02 | |
| *** acormier has joined #openstack-ansible | 01:02 | |
| *** jrobinson has joined #openstack-ansible | 01:03 | |
| *** acormier has quit IRC | 01:08 | |
| *** weezS has joined #openstack-ansible | 01:19 | |
| *** schwicht has joined #openstack-ansible | 01:22 | |
| *** galstrom_zzz is now known as galstrom | 01:26 | |
| *** SerenaFeng has quit IRC | 01:26 | |
| *** SerenaFeng has joined #openstack-ansible | 01:37 | |
| *** schwicht has quit IRC | 01:44 | |
| *** schwicht has joined #openstack-ansible | 01:45 | |
| *** schwicht_at_work has joined #openstack-ansible | 01:47 | |
| *** schwicht has quit IRC | 01:50 | |
| *** Jeffrey4l has joined #openstack-ansible | 01:53 | |
| *** jamielennox is now known as jamielennox|away | 01:56 | |
| *** jamielennox|away is now known as jamielennox | 02:01 | |
| *** eki__ has joined #openstack-ansible | 02:02 | |
| eki__ | someone interested in trying to help me figure out why I can't ping my openStack routers from my aio host? | 02:04 |
| *** schwicht_at_work has quit IRC | 02:12 | |
| *** jwitko has quit IRC | 02:13 | |
| *** Jeffrey4l has quit IRC | 02:14 | |
| *** Jeffrey4l has joined #openstack-ansible | 02:14 | |
| *** cuongnv has quit IRC | 02:29 | |
| *** cuongnv has joined #openstack-ansible | 02:32 | |
| *** sanfern has joined #openstack-ansible | 02:33 | |
| *** winggundamth has joined #openstack-ansible | 02:37 | |
| *** sanfern has quit IRC | 02:39 | |
| *** Jeffrey4l has quit IRC | 02:41 | |
| *** Jeffrey4l has joined #openstack-ansible | 02:41 | |
| *** cathrich_ has joined #openstack-ansible | 02:47 | |
| *** cathrichardson has quit IRC | 02:47 | |
| *** Mahe has quit IRC | 02:48 | |
| *** Mahe has joined #openstack-ansible | 02:50 | |
| *** gouthamr has quit IRC | 03:01 | |
| *** weezS has quit IRC | 03:02 | |
| *** acormier has joined #openstack-ansible | 03:04 | |
| *** weezS has joined #openstack-ansible | 03:04 | |
| *** jrobinson has quit IRC | 03:08 | |
| *** galstrom is now known as galstrom_zzz | 03:10 | |
| *** crushil has joined #openstack-ansible | 03:20 | |
| *** weezS has quit IRC | 03:22 | |
| *** SerenaFeng has quit IRC | 03:24 | |
| *** SerenaFeng has joined #openstack-ansible | 03:24 | |
| *** acormier has quit IRC | 03:27 | |
| *** galstrom_zzz is now known as galstrom | 03:28 | |
| *** SerenaFeng has quit IRC | 03:29 | |
| *** LinStatSDR has joined #openstack-ansible | 03:29 | |
| *** weezS has joined #openstack-ansible | 03:33 | |
| *** markvoelker_ has quit IRC | 03:53 | |
| *** udesale has joined #openstack-ansible | 04:00 | |
| *** LinStatSDR has quit IRC | 04:05 | |
| *** galstrom is now known as galstrom_zzz | 04:07 | |
| openstackgerrit | Ravi Kumar Boyapati proposed openstack/openstack-ansible-rsyslog_client master: Fix remote logging template https://review.openstack.org/446838 | 04:14 |
| eki__ | I am just going to write the problem, if anyone has any advice it would be awesome. | 05:06 |
| eki__ | I am installing openstack aoi on a machine with single interface (eth0). The installation works perfectly, everything seems fine. I can create instances, networks,routers and access horizon. The problem is I can't figure out how to configure the routers so the instances can access internet. | 05:06 |
| eki__ | What should I configure as external network? Is it the br-vxlan or br-vlan networks or my real physical network? There must be few ways to do this, but what would be the correct way of doing this with aoi, or does it simply depend on what I have configured in openstack_user_config.yml, or can I use the default configuration? | 05:06 |
| eki__ | I have been reading the documentation for a bit but can't figure it out | 05:08 |
| eki__ | the main problem seems to be that I can't connect to the routers I create | 05:10 |
| *** sanfern has joined #openstack-ansible | 05:12 | |
| *** crushil has quit IRC | 05:12 | |
| *** Jack_Iv has joined #openstack-ansible | 05:24 | |
| *** sanfern has quit IRC | 05:28 | |
| *** sanfern has joined #openstack-ansible | 05:28 | |
| xdfil | eki__ what is your eth0 config ? | 05:31 |
| eki__ | first tried with static ip | 05:32 |
| xdfil | can i see it? | 05:32 |
| eki__ | paste here? | 05:33 |
| eki__ | just nuked the whole install too | 05:33 |
| xdfil | idk if that will work | 05:33 |
| xdfil | ppl use those paste services | 05:33 |
| *** shausy has joined #openstack-ansible | 05:34 | |
| eki__ | http://pastebin.com/NFRgVEtp | 05:34 |
| xdfil | logan- I fixed the storwize issue, been working on bgp | 05:34 |
| eki__ | that is what I have now since I nuked | 05:34 |
| eki__ | I did have promisc + broadcast before too | 05:34 |
| xdfil | logan- trying to figure out how to make the container bgp iface ip static | 05:35 |
| eki__ | I also tried moving the ip settings to a bridge but I currently don't have real understing what I am supposed to do | 05:35 |
| eki__ | is it supposed to work as it is on clean install? | 05:36 |
| *** NikhilS has joined #openstack-ansible | 05:36 | |
| xdfil | eki__ make a bridge interface and put the IP from eth0 on the bridge interface instead | 05:36 |
| eki__ | will do | 05:36 |
| xdfil | then you should be able to specify the bridge interface as a flat network in neutron | 05:37 |
| eki__ | should I name the bridge to something that openstack expects or separete? | 05:37 |
| xdfil | br-provider is common | 05:37 |
| xdfil | I hate all the bridge names in openstack | 05:39 |
| xdfil | i hate wahat the openstack comunity names things | 05:39 |
| xdfil | eki__ full disclosure: ive never done an AIO install and I've given people bad advice in the past | 05:41 |
| eki__ | ok | 05:41 |
| xdfil | I don't know what an AIO looks like | 05:41 |
| eki__ | should be fine :D | 05:41 |
| eki__ | http://pastebin.com/4b72fktL | 05:41 |
| xdfil | i'm already confused by the fact you have no brdiges for trhe containers? | 05:41 |
| eki__ | I have been hitting my head to wall for a day now | 05:41 |
| eki__ | xdfil: there will be once I run the installer | 05:42 |
| eki__ | "nuked the whole thing, and in process to build it from start" | 05:42 |
| eki__ | sorry shoudl have specified that | 05:42 |
| xdfil | ohhhh | 05:42 |
| eki__ | just running the bootstrap-aio.sh | 05:44 |
| eki__ | have to take look at the openstack_user_config.ym before running the all the other playbooks | 05:45 |
| eki__ | xdfil: thanks for helping out :D | 05:46 |
| *** weezS has quit IRC | 05:48 | |
| eki__ | http://pastebin.com/feZ7Z25x that is the default openstack_user_config.yml with aio | 05:48 |
| eki__ | or what was genereated for me | 05:48 |
| eki__ | there is no mention of eth0 nor br-provider, should there be? | 05:49 |
| eki__ | and state of my interfaces http://pastebin.com/37R15JGm | 05:51 |
| eki__ | in my /etc/network/interfaces.d/osa_interfaces.cfg <- automatically generated (http://pastebin.com/pUQUwVmG) there is this line | 05:55 |
| eki__ | " # To provide internet connectivity to instances" | 05:55 |
| eki__ | which is on br-vxlan | 05:55 |
| eki__ | makes me think I am supposed to configure router with external interface on the same network as the br-vxlan ( in my case 172.29.240/22) | 05:56 |
| eki__ | aand those exact rules are configure for eth0 | 05:57 |
| eki__ | shoudln't those be br-provider | 05:57 |
| *** SerenaFeng has joined #openstack-ansible | 05:57 | |
| eki__ | will give this idea a try | 05:58 |
| xdfil | just do br-mgmt and br-vlan | 05:58 |
| xdfil | but the "flat" onew | 05:59 |
| xdfil | but the "flat" one | 05:59 |
| xdfil | not the "vlan" one | 05:59 |
| eki__ | remove others? | 05:59 |
| xdfil | yes | 05:59 |
| xdfil | you don't need vxlan storage | 06:00 |
| *** SerenaFeng has quit IRC | 06:00 | |
| xdfil | actually | 06:00 |
| xdfil | keep both vlan bridges | 06:00 |
| eki__ | ok | 06:00 |
| eki__ | so no removing of bridges | 06:01 |
| eki__ | or just keep the vlan ones | 06:01 |
| eki__ | aa | 06:01 |
| xdfil | i think i see whats going on they share out the br-vlan as vlan 1 to do a provider network | 06:01 |
| eki__ | brmgmt and br-vlan and br-vlanx | 06:01 |
| xdfil | vlans and mgmt | 06:02 |
| xdfil | no vlanx | 06:02 |
| xdfil | well wait | 06:02 |
| eki__ | *vxlan | 06:02 |
| xdfil | it doesnt mattee | 06:02 |
| *** Jeffrey4l has quit IRC | 06:02 | |
| xdfil | if you want to play with vlanx do it | 06:02 |
| xdfil | but you dont need it to do the basic stuff | 06:03 |
| eki__ | for now I want to figure out how to connect to instance from the host | 06:03 |
| xdfil | like connect instances to the internet | 06:03 |
| xdfil | ogh | 06:03 |
| eki__ | and internet | 06:03 |
| xdfil | connecting to the host can be anissue | 06:03 |
| xdfil | It might not be possible | 06:04 |
| eki__ | hmmm | 06:04 |
| eki__ | connecting from the host --> instance? | 06:04 |
| *** SerenaFeng has joined #openstack-ansible | 06:04 | |
| xdfil | yes its a common issue in virtulization | 06:05 |
| eki__ | how would you connect to it then? | 06:05 |
| eki__ | from other computer in external network? | 06:05 |
| xdfil | yes short answer | 06:05 |
| *** jascott1- is now known as jascott1 | 06:06 | |
| xdfil | idk i'm sure if you messed with it it could work, I dont thing the linux bridge is the issue but maybe LXC networking stuff or apparmor would get in your way | 06:06 |
| xdfil | maybe what i;m thinking only applys to macvtap | 06:07 |
| eki__ | hmmm | 06:07 |
| eki__ | in short no removing of stuff? | 06:09 |
| eki__ | I will build with the defaults , and this time make the router have ip from external group and try pingin from different host | 06:13 |
| eki__ | didn't think that would be a problem | 06:13 |
| *** Jeffrey4l has joined #openstack-ansible | 06:17 | |
| *** Amit82 has joined #openstack-ansible | 06:34 | |
| Amit82 | Hi All | 06:36 |
| eki__ | hi | 06:36 |
| Amit82 | muxdeamon: Yesterday, you shared your openstack_user_config.yml and /etc/network/interfaces files with me. | 06:37 |
| Amit82 | could you please tell that in "external_lb_vip_address: openstack.willsher.systems" openstack.willsher.systems corresponds to IP on which interface | 06:38 |
| Amit82 | ? | 06:38 |
| *** sanfern has quit IRC | 06:38 | |
| Amit82 | does it bind to ip on en0? | 06:38 |
| *** sanfern has joined #openstack-ansible | 06:39 | |
| *** yolanda has joined #openstack-ansible | 06:45 | |
| *** McMurlock1 has joined #openstack-ansible | 06:54 | |
| *** fxpester has joined #openstack-ansible | 07:03 | |
| *** basilAB_ has quit IRC | 07:18 | |
| *** SerenaFeng has quit IRC | 07:19 | |
| *** aetaric has quit IRC | 07:19 | |
| *** mrhillsman has quit IRC | 07:20 | |
| *** vaishali has joined #openstack-ansible | 07:23 | |
| *** basilAB has quit IRC | 07:23 | |
| *** aetaric has joined #openstack-ansible | 07:24 | |
| *** basilAB has joined #openstack-ansible | 07:26 | |
| *** SerenaFeng has joined #openstack-ansible | 07:30 | |
| *** mrhillsman has joined #openstack-ansible | 07:33 | |
| *** manheim has joined #openstack-ansible | 07:35 | |
| *** manheim has quit IRC | 07:37 | |
| *** manheim has joined #openstack-ansible | 07:38 | |
| *** manheim has joined #openstack-ansible | 07:38 | |
| *** pcaruana has joined #openstack-ansible | 07:43 | |
| *** Jack_Iv has quit IRC | 07:52 | |
| *** Jack_Iv has joined #openstack-ansible | 07:52 | |
| *** Jack_Iv has quit IRC | 07:56 | |
| *** sunjon_ has quit IRC | 07:58 | |
| *** Jack_Iv has joined #openstack-ansible | 08:01 | |
| *** sanfern has quit IRC | 08:03 | |
| *** pmannidi has quit IRC | 08:04 | |
| *** sanfern has joined #openstack-ansible | 08:04 | |
| *** DimGR has joined #openstack-ansible | 08:06 | |
| *** Jack_Iv has quit IRC | 08:22 | |
| *** sanfern has quit IRC | 08:23 | |
| *** Jack_Iv has joined #openstack-ansible | 08:23 | |
| *** Jack_Iv has quit IRC | 08:27 | |
| *** shausy has quit IRC | 08:33 | |
| *** shausy has joined #openstack-ansible | 08:34 | |
| *** udesale has quit IRC | 08:41 | |
| *** vnogin has joined #openstack-ansible | 08:41 | |
| *** udesale has joined #openstack-ansible | 08:41 | |
| admin0 | morning all \o | 08:45 |
| eki__ | morning | 08:54 |
| eki__ | xdfil: no luck pinging from outside | 09:00 |
| eki__ | :/ | 09:00 |
| *** shardy has joined #openstack-ansible | 09:01 | |
| eki__ | is there some security/firewall settings that need to be changed to be able to ping routers? | 09:02 |
| *** sanfern has joined #openstack-ansible | 09:05 | |
| *** david-lyle_ has joined #openstack-ansible | 09:18 | |
| *** david-lyle has quit IRC | 09:18 | |
| admin0 | check routing :) | 09:20 |
| *** lwiecek has joined #openstack-ansible | 09:20 | |
| *** karimb has joined #openstack-ansible | 09:24 | |
| *** qiliang27 has quit IRC | 09:33 | |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible stable/mitaka: Update all SHAs for 13.3.17 https://review.openstack.org/446468 | 09:38 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible master: Update all SHAs for Pike 2017-03-16 https://review.openstack.org/446397 | 09:40 |
| *** hamza has joined #openstack-ansible | 09:43 | |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible stable/ocata: Revert to setuptools 33.1.1 https://review.openstack.org/446922 | 09:44 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible stable/newton: Revert to setuptools 33.1.1 https://review.openstack.org/446923 | 09:45 |
| odyssey4me | andymccr unfortunately the new setuptools just got blocked too - I guess we should hold off those changes indefinitely until we see a month of no changes upstream | 09:45 |
| andymccr | odyssey4me: sounds good | 09:46 |
| openstackgerrit | Merged openstack/openstack-ansible-repo_build stable/newton: Updates to support ignored packages and external indexes https://review.openstack.org/446719 | 09:48 |
| odyssey4me | andymccr argh: http://logs.openstack.org/periodic/periodic-openstack-ansible-upgrade-aio-master-ubuntu-xenial/573a6b7/console.html#_2017-03-17_08_06_22_464572 | 09:48 |
| odyssey4me | the swift deployment failed in the upgrade due to the min_part_hours | 09:48 |
| andymccr | odyssey4me: there is a var you can set to get around that - we do that in the swift gate | 09:48 |
| odyssey4me | you've worked around that in the role tests right? all we do is add an override var for the AIO config I guess? | 09:48 |
| andymccr | swift_pretend_min_part_hours_passed: True | 09:49 |
| DimGR | when doing in Ocata 15.0 lxc-container-create.yml --limit neutron_all it complains about http://paste.openstack.org/show/603020/ if i run it without the --limit it finishes fine . is this a user error or bug ? | 09:51 |
| *** SerenaFeng has quit IRC | 09:57 | |
| *** vnogin has quit IRC | 10:03 | |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible master: Bypass min_part_hours check for the AIO https://review.openstack.org/446932 | 10:04 |
| odyssey4me | andymccr ^ | 10:04 |
| andymccr | odyssey4me: btw do you still have the failed gate? i think there may be a bug there that can be fixed | 10:05 |
| odyssey4me | andymccr what do you mean? | 10:05 |
| odyssey4me | DimGR if you want to limit for that playbook you must also include lxc_hosts as the playbook uses host delegation and needs to gather facts from them | 10:06 |
| odyssey4me | ie lxc-container-create.yml --limit lxc_hosts,neutron_all | 10:06 |
| andymccr | odyssey4me: well there should be no change in the ring so it shouldnt try a rebalance, i think it tries a rebalance and doesnt recognize that there is no change (possibly) | 10:07 |
| odyssey4me | andymccr hmm, this is after an upgrade so surely it should need to anyway due to the new code? | 10:08 |
| andymccr | odyssey4me: the ring shouldnt change if the hosts/disks haven't afaik, but yeah i'll take a look - i can recreate that pretty easily | 10:08 |
| odyssey4me | andymccr that set of task doesn't seem to execute on any conditional except for being the primary in the ring: https://github.com/openstack/openstack-ansible-os_swift/blob/master/tasks/swift_rings.yml#L26 | 10:10 |
| odyssey4me | https://github.com/openstack/openstack-ansible-os_swift/blob/master/tasks/swift_rings_build.yml#L46 | 10:10 |
| odyssey4me | there's no conditional to stop it building | 10:11 |
| odyssey4me | I'm guessing that you might have meant for the swift_rings_check to produce a result, then for the build to happen if that result was that it was needed? | 10:12 |
| *** cuongnv has quit IRC | 10:15 | |
| *** shausy has quit IRC | 10:16 | |
| *** shausy has joined #openstack-ansible | 10:16 | |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible stable/newton: Updates to support ignored packages and external indexes https://review.openstack.org/446718 | 10:18 |
| andymccr | odyssey4me: yeah it'll happen in the execution of the script itself (so swift_rings.py shouldnt try the rebalance so it wouldnt then fail) but that can be fixed :) | 10:20 |
| odyssey4me | ah | 10:21 |
| odyssey4me | I would think that it'd be better if the script exited with a different return code instead of failing in that condition. | 10:21 |
| odyssey4me | ie 0 = success, 1 = fail, 2 = did nothing | 10:22 |
| odyssey4me | or something like that | 10:22 |
| andymccr | odyssey4me: yeah i thinkso - although i think the problem is we call the ringbuilder swift bits directly now and that will error 1 if you try rebalance and it cant rebalance | 10:23 |
| andymccr | so we will need some logic in there but it should be easy enough | 10:23 |
| odyssey4me | the alternative would be to add a '--check' parameter which is read only and informs the automation what needs doing | 10:23 |
| andymccr | thats a new change in Pike (for our swift role) so it should be fine in Ocata/Newton still | 10:23 |
| odyssey4me | perhaps we can get something like that built into the swift tooling so that everyone can benefit from it | 10:24 |
| odyssey4me | (similar to what we arranged with keystone) | 10:24 |
| *** deadnull has joined #openstack-ansible | 10:30 | |
| *** deadnull has quit IRC | 10:30 | |
| odyssey4me | bbiab - off to run some errands | 10:32 |
| andymccr | i need to run some errands this afternoon so i probably wont be back online today (or at least much later). | 10:37 |
| *** udesale has quit IRC | 10:48 | |
| *** vnogin has joined #openstack-ansible | 10:51 | |
| *** smatzek has joined #openstack-ansible | 10:52 | |
| *** vnogin has quit IRC | 10:52 | |
| *** vnogin has joined #openstack-ansible | 10:52 | |
| *** Andrew_jedi has joined #openstack-ansible | 10:56 | |
| *** rcarrillocruz has quit IRC | 11:00 | |
| *** karimb has quit IRC | 11:04 | |
| *** karimb has joined #openstack-ansible | 11:06 | |
| *** hamza has quit IRC | 11:38 | |
| *** Jack_Iv has joined #openstack-ansible | 11:40 | |
| *** Jack_Iv has quit IRC | 11:40 | |
| *** Jack_Iv has joined #openstack-ansible | 11:40 | |
| *** winggundamth has quit IRC | 11:40 | |
| *** shausy has quit IRC | 11:51 | |
| *** shausy has joined #openstack-ansible | 11:51 | |
| bww | Hi Amit82: Any luck on AIO external connectivity? | 11:51 |
| *** shardy is now known as shardy_lunch | 12:04 | |
| *** Andrew_jedi has quit IRC | 12:07 | |
| *** Andrew_jedi has joined #openstack-ansible | 12:12 | |
| *** LinStatSDR has joined #openstack-ansible | 12:14 | |
| *** Jack_Iv_ has joined #openstack-ansible | 12:15 | |
| Amit82 | bww: we are able to have external connectivity | 12:17 |
| Amit82 | but we are not using AIO | 12:17 |
| Amit82 | we are having two separate machines acting as Controller and Compute node respectively | 12:18 |
| *** Jack_Iv has quit IRC | 12:18 | |
| bww | ok, can you share your deployment configuration and any guides on how you deplyed? | 12:20 |
| bww | deployed | 12:20 |
| Amit82 | I followed the the newton deployment guide | 12:20 |
| Amit82 | as I have installed Newton release using 14.0.8 tag | 12:21 |
| bww | ok, did you have to make any modifications to networking? | 12:21 |
| *** LinStatSDR has left #openstack-ansible | 12:21 | |
| *** woodard has quit IRC | 12:24 | |
| *** Oku_OS is now known as Oku_OS-away | 12:24 | |
| *** Oku_OS-away is now known as Oku_OS | 12:24 | |
| *** woodard has joined #openstack-ansible | 12:24 | |
| Amit82 | bww: I have pasted /etc/network/interfaces from both the nodes and openstack_user_config.yml here: http://paste.openstack.org/show/603118/ | 12:26 |
| Amit82 | hope it helps | 12:26 |
| Amit82 | We are still facing some issue of loosing connectivity to Compute and Controller node, if we use two NICs | 12:27 |
| Amit82 | but if we use one extra NIC (eth2) in addition to eth1, for accessing lab n/w, things are working fine | 12:28 |
| bww | cool thanks...I am fairly new to deployiong Openstack via Ansible, but getting the hang of it | 12:29 |
| *** Jack_Iv_ has quit IRC | 12:33 | |
| *** markvoelker has joined #openstack-ansible | 12:35 | |
| *** fandi has joined #openstack-ansible | 12:39 | |
| *** rpittau has quit IRC | 12:41 | |
| bww | Amit82 | 12:45 |
| bww | Is your setup with two servers and the nova is a physical server while the rest of the services like storage, etc are runing on the controller node as LXC containers? | 12:46 |
| *** shausy has quit IRC | 12:48 | |
| Amit82 | bww: I am having this example setup: https://docs.openstack.org/project-deploy-guide/openstack-ansible/newton/app-config-test.html#test-environment-config | 12:50 |
| Amit82 | "Block Storage Host" is not there in my case | 12:50 |
| bww | ok cool, so your not running a seprate storage node | 12:52 |
| *** Jack_Iv has joined #openstack-ansible | 12:52 | |
| *** acormier has joined #openstack-ansible | 12:52 | |
| *** acormier has quit IRC | 12:52 | |
| *** acormier has joined #openstack-ansible | 12:52 | |
| *** schwicht has joined #openstack-ansible | 12:54 | |
| mgariepy | cloudnull, 3.10.0-514.2.2.el7.x86_64 | 12:54 |
| mgariepy | cloudnull, starting with this image, yum update -y, https://cloud.centos.org/centos/7/images/CentOS-7-x86_64-GenericCloud-1612.raw.tar.gz | 12:55 |
| *** acormier has quit IRC | 12:55 | |
| *** Amit82 has quit IRC | 12:55 | |
| *** acormier has joined #openstack-ansible | 12:55 | |
| *** askb has quit IRC | 12:56 | |
| *** manheim has quit IRC | 12:56 | |
| *** manheim has joined #openstack-ansible | 12:56 | |
| mgariepy | cloudnull, http://paste.openstack.org/show/603127/ | 12:57 |
| *** acormier has quit IRC | 12:57 | |
| *** muxdaemon has quit IRC | 12:59 | |
| *** schwicht has quit IRC | 13:03 | |
| *** schwicht has joined #openstack-ansible | 13:05 | |
| *** hamza has joined #openstack-ansible | 13:06 | |
| *** karimb has quit IRC | 13:09 | |
| *** esberglu has joined #openstack-ansible | 13:13 | |
| *** esberglu has left #openstack-ansible | 13:13 | |
| *** esberglu has joined #openstack-ansible | 13:13 | |
| *** karimb has joined #openstack-ansible | 13:18 | |
| openstackgerrit | Merged openstack/openstack-ansible stable/newton: Revert to setuptools 33.1.1 https://review.openstack.org/446923 | 13:18 |
| *** schwicht has quit IRC | 13:21 | |
| *** gouthamr has joined #openstack-ansible | 13:23 | |
| *** muxdaemon has joined #openstack-ansible | 13:25 | |
| *** hamza has quit IRC | 13:29 | |
| Andrew_jedi | mgariepy: Hi, quick question, If i want to look for neutron code in CentOS7 then i should like in "/usr/share/nova" directory ? | 13:30 |
| Andrew_jedi | s/like/look | 13:30 |
| *** muxdaemon has quit IRC | 13:32 | |
| *** smatzek has quit IRC | 13:32 | |
| odyssey4me | Andrew_jedi our deployment on CentOS is the same as on Ubuntu for the most part | 13:33 |
| odyssey4me | certainly for the venvs | 13:33 |
| odyssey4me | ie you'll find the venvs in /openstack/venvs | 13:33 |
| *** cathrichardson has joined #openstack-ansible | 13:34 | |
| *** LinStatSDR has joined #openstack-ansible | 13:34 | |
| Andrew_jedi | odyssey4me: Ohh, yes, i forgot about the venvs. Actually this particular installation is done via ansible but not using OSA for now. I was looking for the location of the actual code to debug an issue. | 13:34 |
| odyssey4me | ah ok | 13:34 |
| *** cathrich_ has quit IRC | 13:36 | |
| openstackgerrit | Merged openstack/openstack-ansible stable/ocata: Revert to setuptools 33.1.1 https://review.openstack.org/446922 | 13:37 |
| admin0 | cannot get telemetry to work on AIO | 13:37 |
| admin0 | :( | 13:37 |
| *** LinStatSDR has left #openstack-ansible | 13:38 | |
| odyssey4me | admin0 do you have a bug report for it? | 13:40 |
| odyssey4me | have you figured out anything particular wrong? | 13:40 |
| admin0 | i am working on gathering all the data | 13:41 |
| odyssey4me | the telemetry stuff is a bit weird | 13:41 |
| Andrew_jedi | admin0: It was working for me, what is the problem that you are facing ? | 13:42 |
| admin0 | i think first issue i face is “how to verify its working” :D | 13:42 |
| Andrew_jedi | admin0: Check metrics via ceilometer client, create an autoscaling template and test alarms | 13:43 |
| admin0 | what API commands would verify that its working .. like i do for openstack server list ; openstack image list ; openstack volume list etc | 13:44 |
| *** cpuga has joined #openstack-ansible | 13:45 | |
| *** schwicht has joined #openstack-ansible | 13:45 | |
| mgariepy | Andrew_jedi, if you are installing from pkg you can list pkg with `rpm -qa` and file with `rpm -qf <pkg>` | 13:47 |
| mgariepy | oops, 'rpm -ql <pkg>' | 13:48 |
| mgariepy | to list files. -qf can tell you which package a file belong to. | 13:48 |
| Andrew_jedi | mgariepy: Awesome, thanks! | 13:48 |
| mgariepy | it's impressive what breaks when you checkout cinder role in place of nova... | 13:52 |
| *** crushil has joined #openstack-ansible | 13:53 | |
| *** FrankZhang has joined #openstack-ansible | 13:53 | |
| *** hamza has joined #openstack-ansible | 13:55 | |
| odyssey4me | lol | 13:55 |
| *** shardy_lunch is now known as shardy | 13:56 | |
| *** smatzek has joined #openstack-ansible | 13:57 | |
| *** woodard has quit IRC | 13:58 | |
| *** FrankZhang has quit IRC | 13:58 | |
| *** FrankZhang has joined #openstack-ansible | 13:58 | |
| Andrew_jedi | mgariepy: just found mitaka code in newton package. :p | 14:03 |
| mgariepy | Andrew_jedi, where the pakcage comes from ? | 14:04 |
| Andrew_jedi | RDO repo, i am looking for a word in english language to express extreme surprise ... | 14:05 |
| *** fabg has joined #openstack-ansible | 14:13 | |
| *** lwiecek has quit IRC | 14:14 | |
| Andrew_jedi | mgariepy: Correction. Packages are fine, somebody decided to modify one of package file, figured it out via "rpm --verify". | 14:16 |
| mgariepy | hehe | 14:16 |
| openstackgerrit | Merged openstack/openstack-ansible-ops master: define the default kernel statically https://review.openstack.org/446790 | 14:16 |
| Andrew_jedi | mgariepy: :p | 14:24 |
| *** cjloader has joined #openstack-ansible | 14:24 | |
| fabg | odyssey4me: hi, did you have time to see https://bugs.launchpad.net/openstack-ansible/+bug/1670632. I know you have more to do :-) Or did someone encounter the same issue on 14.1.1 ? | 14:24 |
| openstack | Launchpad bug 1670632 in openstack-ansible "ceilometer error because gnocchiclient > 3.0 for stable/newton " [Undecided,New] - Assigned to Jesse Pretorius (jesse-pretorius) | 14:24 |
| admin0 | Andrew_jedi:….. for a word in english language to express extreme surprise => Massively surprized | 14:25 |
| Andrew_jedi | admin0: try "flabbergasted". | 14:26 |
| admin0 | and i thought it was german word :D | 14:26 |
| admin0 | or nordic | 14:26 |
| odyssey4me | fabg yeah, sorry - I've been trying to confirm it but have been side-tracked with other issues | 14:27 |
| *** chris_hultin|AWA is now known as chris_hultin | 14:30 | |
| fabg | odyssey4me: i understand ;-) I follow the osa community activity and i see the mass of work you slaughter :-) | 14:30 |
| *** foutatoro has joined #openstack-ansible | 14:33 | |
| foutatoro | hi, I try to ping an instance in project network from the namespace of dhcp but I get the error "ping: error while loading shared libraries: libcap.so.2: cannot stat shared object: Permission denied" | 14:35 |
| foutatoro | did someboby face in OSA with issue before ? | 14:36 |
| *** rpittau has joined #openstack-ansible | 14:38 | |
| *** cathrichardson has quit IRC | 14:39 | |
| *** cathrichardson has joined #openstack-ansible | 14:39 | |
| *** jmckind has joined #openstack-ansible | 14:42 | |
| *** agrebennikov has joined #openstack-ansible | 14:46 | |
| *** sc68cal has quit IRC | 14:47 | |
| *** kstev has joined #openstack-ansible | 14:48 | |
| *** vnogin has quit IRC | 14:48 | |
| *** acormier has joined #openstack-ansible | 14:50 | |
| *** fandi has quit IRC | 14:53 | |
| *** looking_around has joined #openstack-ansible | 14:54 | |
| *** looking_around has left #openstack-ansible | 14:55 | |
| *** Dinesh_Bhor has quit IRC | 14:56 | |
| *** manheim has quit IRC | 14:57 | |
| *** marst has quit IRC | 14:58 | |
| foutatoro | any suggestion ? is it a bug with xenial ? .. | 14:58 |
| strattao | for the provider networks specified in the openstack_user_config.yml, do I have to specify a “flat” network? | 15:03 |
| *** marst has joined #openstack-ansible | 15:04 | |
| strattao | or can I just use only a vlan provider network for the br-vlan container bridge connection? | 15:04 |
| *** ansibleRhino has joined #openstack-ansible | 15:04 | |
| *** marst has quit IRC | 15:05 | |
| admin0 | strattao . is it post-install question ? | 15:05 |
| *** marst has joined #openstack-ansible | 15:05 | |
| *** fxpester has quit IRC | 15:05 | |
| strattao | yes - I get an error saying that the flat network is not defined | 15:05 |
| strattao | but I don’t want a flat network… | 15:05 |
| *** aludwar has quit IRC | 15:05 | |
| *** aludwar has joined #openstack-ansible | 15:06 | |
| *** sanfern has quit IRC | 15:06 | |
| strattao | and can’t figure out what is expecting to use a flat network | 15:06 |
| *** fabg has quit IRC | 15:07 | |
| *** sanfern has joined #openstack-ansible | 15:07 | |
| *** messy has quit IRC | 15:08 | |
| *** NikhilS has quit IRC | 15:08 | |
| *** hamza has quit IRC | 15:10 | |
| admin0 | strattao: you don’t have to use flat network .. you can use vlan to add exsternal network — http://www.openstackfaq.com/openstack-add-floating-ips/ | 15:12 |
| *** manheim has joined #openstack-ansible | 15:15 | |
| *** Jack_Iv has quit IRC | 15:20 | |
| *** Jack_Iv has joined #openstack-ansible | 15:21 | |
| *** manheim has quit IRC | 15:22 | |
| *** acormier has quit IRC | 15:28 | |
| *** vnogin has joined #openstack-ansible | 15:30 | |
| *** foutatoro has quit IRC | 15:31 | |
| *** vnogin has quit IRC | 15:34 | |
| *** vnogin has joined #openstack-ansible | 15:34 | |
| openstackgerrit | Tom Jose Kalapura proposed openstack/openstack-ansible-rsyslog_client stable/newton: Fix remote logging template https://review.openstack.org/447036 | 15:39 |
| agrebennikov | seems st patrick's day killed the channel for today :D | 15:41 |
| spotz | agrebennikov: It's not helping it no:) | 15:42 |
| spotz | I think our fearless leader may still be in Milan or travelling back | 15:43 |
| *** galstrom_zzz is now known as galstrom | 15:43 | |
| agrebennikov | but usually there is pretty active discussion happening by this time during the week... even on fridays ;) | 15:44 |
| spotz | Just blame St. Paddy's:) | 15:44 |
| *** crushil has quit IRC | 15:45 | |
| *** Jack_Iv has quit IRC | 15:49 | |
| strattao | thanks admin0 | 15:50 |
| *** hamza has joined #openstack-ansible | 16:02 | |
| *** xinli has joined #openstack-ansible | 16:02 | |
| *** vishwanathj has joined #openstack-ansible | 16:03 | |
| admin0 | strattao: were yu able to add the IPs ? | 16:04 |
| *** Oku_OS is now known as Oku_OS-away | 16:05 | |
| *** chris_hultin is now known as chris_hultin|AWA | 16:07 | |
| *** muxdaemon has joined #openstack-ansible | 16:12 | |
| *** MasterOfBugs has joined #openstack-ansible | 16:17 | |
| *** vishwanathj has quit IRC | 16:17 | |
| *** shananigans has quit IRC | 16:23 | |
| *** zz_pwnall1337 is now known as pwnall1337 | 16:25 | |
| *** marst has quit IRC | 16:28 | |
| *** muxdaemo_ has joined #openstack-ansible | 16:33 | |
| *** Andrew_jedi has quit IRC | 16:36 | |
| *** muxdaemon has quit IRC | 16:37 | |
| *** muxdaemo_ has quit IRC | 16:38 | |
| *** muxdaemon has joined #openstack-ansible | 16:39 | |
| *** crushil has joined #openstack-ansible | 16:40 | |
| *** cmart has joined #openstack-ansible | 16:42 | |
| *** Jack_Iv has joined #openstack-ansible | 16:50 | |
| *** ansibleRhino has left #openstack-ansible | 16:50 | |
| *** vnogin has quit IRC | 16:52 | |
| *** Jack_Iv has quit IRC | 16:53 | |
| *** acormier has joined #openstack-ansible | 16:57 | |
| *** Andrew_jedi has joined #openstack-ansible | 16:59 | |
| cmart | is anyone here using nova libvirt password/key injection? it seems that libguestfs may be broken for OSA Newton. | 17:05 |
| odyssey4me | csmart not as far as I know? although I'd like to understand how it's an OSA issue and not a nova/libguestfs issue? | 17:07 |
| odyssey4me | perhaps a config problem? | 17:07 |
| odyssey4me | by that I mean nova.conf? | 17:07 |
| csmart | odyssey4me: cmart ^ | 17:08 |
| cmart | odyssey4me, nova.conf looks good, I have "inject_partition = -1", "inject_password = True", "inject_key = True" | 17:08 |
| csmart | Man, our nics are a bit too similar... | 17:09 |
| cmart | when I launch an instance I get the following in nova-compute.log: "Ignoring error injecting data into image <LocalFileImage:{'path': '/var/lib/nova/instances/0c32099c-c6da-4505-a8a0-69bd26b2bb5f/disk', 'format': 'qcow2'}> (libguestfs installed but not usable (/usr/bin/supermin exited with error status 1." | 17:09 |
| odyssey4me | lol | 17:09 |
| * csmart goes back to sleep :-) | 17:09 | |
| csmart | \o | 17:09 |
| odyssey4me | hmm, tell me that whatever is trying to do this is running on bare metal - not in a container? | 17:10 |
| odyssey4me | whoops, sorry csmart | 17:10 |
| *** hamza has quit IRC | 17:10 | |
| cmart | odyseey4me, yes, this is on a compute host | 17:10 |
| odyssey4me | as far as I recall, that function should be happening on the compute hosts - and libguestfs can only build the supermin on a bare metal host | 17:10 |
| odyssey4me | hmm, ok, so either we're missing a package, some config or something is broken in ubuntu? | 17:11 |
| odyssey4me | can you validate with someone who has done this at all that you have the right config in nova.conf and the right packages - it might be worth pinging the openstack-operators list about it | 17:12 |
| csmart | odyssey4me: all good, even I do a double take when cmart talks :-) later | 17:12 |
| odyssey4me | of course, one could hope that someone in here has a working setup and can help troubleshoot | 17:13 |
| odyssey4me | not sure if logan- or jmccrory make use of that and can help troubleshoot | 17:13 |
| odyssey4me | or agrebennikov | 17:14 |
| cmart | odyssey4me could do, but the libvirt section of nova.conf looks good according to the various blog entries I've read. I believe this is the code that Nova executes to produce the error: https://github.com/openstack/nova/blob/c6cb5cf1ba0b1483951f3d236c53ff7924f3d7af/nova/virt/disk/vfs/guestfs.py#L76 | 17:15 |
| cmart | Also, below that error is this: "To see full error messages you may need to enable debugging. Do: export LIBGUESTFS_DEBUG=1 LIBGUESTFS_TRACE=1 and run the command again." | 17:16 |
| cmart | (I'm not sure how to pass environment variables to the running Nova service) | 17:17 |
| cmart | So I'm unsure if this is an OSA issue at all, just curious if others have gotten this working. | 17:17 |
| *** pcaruana has quit IRC | 17:18 | |
| cmart | overall I'd rather not do it this way and just use cloud-init to inject the SSH key, but we have some years-old images without cloud-init, that we still need to support. | 17:23 |
| odyssey4me | yeah, I think the majority of people use cloud-init | 17:26 |
| agrebennikov | cmart, sorry, I guess I'm missing the context a little bit... how may I help having baremetals? | 17:28 |
| *** muxdaemon has quit IRC | 17:29 | |
| agrebennikov | odyssey4me, is that still valid? | 17:29 |
| cmart | agrebennikov just having a trouble with libvirt key injection on the compute host, which uses libguestfs. | 17:30 |
| agrebennikov | config drive you mean? | 17:30 |
| odyssey4me | no, the libvirt key injection actually modifies the image on the compute host using libguestfs instead of using cloud-init (which uses config drive) IIRC | 17:31 |
| agrebennikov | ah, yeah, sorry... so... we use cloud-init everywhere, but if I only need to change one compute node configuration - I can test it out | 17:32 |
| *** woodard has joined #openstack-ansible | 17:34 | |
| cmart | agrebennikov, only if you feel like trying it out. i'm also continuing to troubleshoot. All I did was set the following three YAML variables: | 17:34 |
| cmart | "nova_libvirt_inject_key: True", "nova_libvirt_inject_partition: -1", "nova_libvirt_inject_password: True" | 17:35 |
| *** chris_hultin|AWA is now known as chris_hultin | 17:35 | |
| cmart | I don't advocate that others go this route if you already use cloud-init and it's working for you! This is only so we can continue supporting some older, cloud-init-less images maintained by our community | 17:36 |
| odyssey4me | freebsd? | 17:37 |
| xdfil | odyssey4me: What's the appropriate way to give a container a static IP? | 17:38 |
| *** chris_hultin is now known as chris_hultin|AWA | 17:38 | |
| cmart | odyssey4me nope, customized builds of older Ubuntu and CentOS, which from a security perspective should be phased out, but people have used them for scientific work | 17:38 |
| *** muxdaemon has joined #openstack-ansible | 17:39 | |
| cmart | odyssey4me for example if you want to try to reproduce a genomics study done 3 years ago, we let you launch the exact same image that was used to run the original analysis, so you can hopefully see how the results were obtained. it's a data provenance thing. | 17:40 |
| odyssey4me | xdfil you don't, unless you want to hack your inventory.json once the inventory item has already been generated | 17:40 |
| *** SerenaFeng has joined #openstack-ansible | 17:41 | |
| xdfil | I have to for BGP speaker. the physical routers need to peer with the IP address of the dragent | 17:41 |
| odyssey4me | cmart ah, yeah - I tried it a few years ago and think I got it right for freebsd | 17:42 |
| odyssey4me | but that was back for grizzly, and I don't have access ot that environment any more | 17:42 |
| *** Jack_Iv has joined #openstack-ansible | 17:42 | |
| xdfil | wont the inventory.json get overwritten later on? | 17:42 |
| odyssey4me | xdfil once an item is generated it will stay there for the lifetime of the container | 17:42 |
| xdfil | ok cool thanks | 17:43 |
| odyssey4me | xdfil if you really need a static IP you can set that service to run on_metal and dedicate a host (or hosts) to it instead of using containers | 17:43 |
| odyssey4me | that's something useful to do with the network_hosts in large environments | 17:43 |
| xdfil | ahh right | 17:44 |
| xdfil | ya know what I like that better | 17:44 |
| xdfil | way better actually i dont need a br-bgp network anymore | 17:45 |
| xdfil | wait... that would mean all the neutron agents would be on metel | 17:45 |
| agrebennikov | so cmart if I do that on one of the computes - there still will be cloud-init workflow in place... how do I know if the stuff you need works? | 17:46 |
| agrebennikov | or you want to give me an image? | 17:46 |
| cmart | agrebennikov it's probably best if you try launching an image with cloud-init disabled or absent. happy to give you one. | 17:50 |
| agrebennikov | sure | 17:50 |
| agrebennikov | can I just manually change the config of nova on one of the computes? | 17:51 |
| agrebennikov | or you want me to do it via osa? | 17:51 |
| *** chris_hultin|AWA is now known as chris_hultin | 17:51 | |
| agrebennikov | (sure - meant to say "sure, give it to me" ;) | 17:51 |
| odyssey4me | xdfil yep, you'd need to dedicate at least two hosts (to allow router migration if one goes down) and all the agents would live on there | 17:51 |
| odyssey4me | agrebennikov I don't think it matters - what cmart needs is a working configuration to compare against. How it got there is immaterial. | 17:52 |
| agrebennikov | that works... | 17:53 |
| odyssey4me | Even if it's some other deployment tool, built using packages or whatever. | 17:53 |
| cmart | yes, I did it via OSA but you may not need to. I think all OSA does is template out the vars in nova.conf. how can I send you a 2 GB qcow2 image? | 17:53 |
| *** shananigans has joined #openstack-ansible | 17:53 | |
| agrebennikov | but I'll only do it if you vote for https://review.openstack.org/#/c/444491/2 and https://review.openstack.org/#/c/425997/18 :D | 17:53 |
| agrebennikov | cmart, you probably can.... | 17:53 |
| agrebennikov | if you for example upload to gdrive | 17:54 |
| agrebennikov | and send me the link | 17:54 |
| cmart | yep, or Amazon S3. give me a few minutes :) | 17:54 |
| *** crushil has quit IRC | 17:55 | |
| *** muxdaemon has quit IRC | 17:55 | |
| odyssey4me | agrebennikov you asked for it ;) | 18:02 |
| odyssey4me | hopefully my review makes sense | 18:02 |
| agrebennikov | so essentially you want me to populate the cert back to the deployment host from the galera cert once it is created? | 18:04 |
| agrebennikov | *from the galera host | 18:04 |
| agrebennikov | odyssey4me, | 18:06 |
| odyssey4me | yeah, everything is massively simpler and more reliable when the deployment host is used as a place to pull to and push from | 18:08 |
| DimGR | it took me 30 days to figure that out ^^ | 18:08 |
| odyssey4me | then even self generated certs are 'user configurated', so there's only one way it ever gets there | 18:08 |
| odyssey4me | we have a bunch of distribution mechanisms that use delegation which we should ideally get rid of and replace with this simpler, more reliable mechanism | 18:09 |
| odyssey4me | ideally every ssl cert generation and placement fits there, and perhaps other things | 18:09 |
| odyssey4me | we're ideally hoping to move towards a place where we don't need a ssh service in the containers - it's far more secure | 18:10 |
| odyssey4me | right now it's only the container to container delegation that's holding that back as far as I know | 18:10 |
| *** cjloader has quit IRC | 18:11 | |
| odyssey4me | in case you're not aware we already connect to containers via the host, rather than directly via ssh | 18:11 |
| *** cjloader has joined #openstack-ansible | 18:11 | |
| xdfil | odyssey4me: I'm thinking of making the container/infra hosts ( i have 3 ) be bare metal for neutron_agents | 18:12 |
| xdfil | you forsee any issues with that? | 18:13 |
| *** cjloader has quit IRC | 18:13 | |
| xdfil | OSA wise | 18:13 |
| odyssey4me | xdfil you mean having your infra hosts with all the containers on them also host the agents, but on bare metal instead of in containers? | 18:13 |
| xdfil | yes | 18:13 |
| *** cjloader has joined #openstack-ansible | 18:13 | |
| odyssey4me | well, it depends on your use-case | 18:13 |
| *** SerenaFeng has quit IRC | 18:14 | |
| odyssey4me | in that situation if something breaks on your network agent hosts, you also lose a third of your api infrastructure when you need to rebuild the hosts | 18:14 |
| odyssey4me | if you're doing that, better to containerise the agents | 18:14 |
| odyssey4me | the containers are easy to replace, the host is not | 18:14 |
| odyssey4me | if it's a seperate host, then you only affect that one smaller set of services if everything goes belly up | 18:15 |
| xdfil | hmmm good points | 18:15 |
| odyssey4me | the whole reason we've designed it to be so flexible in terms of service-host allocation is so that you really can spread and scale easily | 18:16 |
| odyssey4me | but you can compress if your budget can't handle it, but that compression should be done with services in containers | 18:16 |
| *** poopcat has joined #openstack-ansible | 18:16 | |
| xdfil | in my case I have 8 nodes that are way too beefy to be anything other than compute hosts | 18:17 |
| xdfil | 3 nodes that are less-beefy more suited for controller work | 18:17 |
| xdfil | I think it makes sense, to do bare metal neutron agents on the 3 controller nodes | 18:19 |
| odyssey4me | sounds like you should stick to the basic 3 controller model then | 18:19 |
| xdfil | I'm just concerned about running into an issue with the OSA runs | 18:19 |
| odyssey4me | well, that's up to you - I'd not advise it operationally, but it's your environment :) | 18:20 |
| odyssey4me | the other concern is security | 18:20 |
| odyssey4me | containers segregate the kernel name spaces for all the services | 18:20 |
| odyssey4me | I don't know what effect it will have if your host is holding those services and all the containers with the possibility of namespace clashes or overlap. | 18:21 |
| xdfil | so if agents running on metal, and the agent gets compromised it can own all the containers is the security concern | 18:21 |
| *** acormier has quit IRC | 18:23 | |
| *** crushil has joined #openstack-ansible | 18:23 | |
| *** poopcat has quit IRC | 18:24 | |
| xdfil | odyssey4me if I configured all the hosts pre-deploy with OVS bridges instead of linux bridges, and specified those bridges in user_settings.yml in place of the br-mgmt/br-vlan would OSA be able to attach containers to them? | 18:26 |
| xdfil | slightly off topic, but I'm curious about that | 18:27 |
| odyssey4me | I honestly have no clue | 18:27 |
| xdfil | :) | 18:27 |
| *** jmckind has quit IRC | 18:28 | |
| *** poopcat has joined #openstack-ansible | 18:32 | |
| xdfil | odyssey4me another thing I am getting hung up on. If an OVS bridge (say br-tun) is connected to a linux bridge ( say br-vxlan ) which interface do I configure the IP on that neutron uses for local_ip | 18:32 |
| odyssey4me | xdfil I honestly have no clue. Networking is something of a dark art to me. | 18:37 |
| xdfil | my network instincts tell me that putting an IP on an interface causes it to become a layer 3 interface and it will nto forward layer 2 frames | 18:37 |
| odyssey4me | I figure things like that out by trying, failing, then trying again. | 18:37 |
| xdfil | odyssey4me: I know I'm just trying to embaras you :) | 18:37 |
| xdfil | thanks though | 18:37 |
| odyssey4me | The best thing you can do is figure out what you think you need, then try it. Then dig into any problems that arise and try to break it and hack it. | 18:38 |
| odyssey4me | Once you've learned, wipe it all and try again. | 18:38 |
| odyssey4me | For a new cloud environment, do this many, many times over. | 18:38 |
| odyssey4me | .For an existing production cloud - leave it and make sure you have a sufficient lab to test with. | 18:39 |
| *** vnogin has joined #openstack-ansible | 18:39 | |
| xdfil | yeah, i'm time constrained. I'm trying to solve problems without stumbling into unexpected dragons | 18:39 |
| odyssey4me | Then once you;ve worked out where you want to be, work out how you get there | 18:39 |
| *** poopcat has quit IRC | 18:40 | |
| odyssey4me | I used to have an old server setup with a bunch of VM's and automation to build, rebuild and do it again and again. | 18:40 |
| odyssey4me | Something similar to the multi-node-aio in the ops repo | 18:40 |
| odyssey4me | It's a vital tool if you're designing/deploying/supporting/maintaining a cloud. | 18:40 |
| *** david-lyle_ is now known as david-lyle | 18:43 | |
| xdfil | odyssey4me I was looking at inventory.json and the containers don't say DHCP they have addresses. So every time they reboot they will get that same address via DHCP? | 18:44 |
| xdfil | but if i destroy/create it will change or stay the same? | 18:44 |
| odyssey4me | xdfil the container eth0 is DHCP and NATted through the host, which is why we don't use them... the addresses you're seeing are the br-mgmt and others which are static for the lifetime of the container | 18:45 |
| odyssey4me | if you destroy/recreate it will stay the same | 18:45 |
| odyssey4me | the only way to change it is to remove it from the inventory, or edit it | 18:45 |
| xdfil | hmmm ok mine changed at some point but I dont rememeber what i did | 18:46 |
| xdfil | thanks | 18:46 |
| odyssey4me | it will only change if you remove it, or remove the inventory entirely | 18:46 |
| *** galstrom is now known as galstrom_zzz | 18:48 | |
| *** poopcat has joined #openstack-ansible | 18:52 | |
| xdfil | odyssey4me I just noticed OSA didn't connect the agents container to the bridge I defined in user-settings | 18:56 |
| *** retreved has joined #openstack-ansible | 18:57 | |
| xdfil | veth get created during lxc-containers-create? | 18:57 |
| odyssey4me | yep | 19:00 |
| *** galstrom_zzz is now known as galstrom | 19:01 | |
| agrebennikov | odyssey4me, can you please make this https://review.openstack.org/#/c/446754/1 happen? I got votes for master, but evrardjp seems to be off today :( | 19:08 |
| *** foutatoro has joined #openstack-ansible | 19:09 | |
| odyssey4me | agrebennikov sure, looks about right - can you add a release note (or edit the existing one) to specify that 'haproxy_bufsize' has been removed in master | 19:11 |
| agrebennikov | oh yeah, sorry about that | 19:11 |
| odyssey4me | it should be an 'upgrade' release note, and should say that it's been removed, and it should be replaced by using the new tuning option | 19:12 |
| agrebennikov | you mean, for backports it will be both "features" and "upgrades"? | 19:13 |
| openstackgerrit | Merged openstack/openstack-ansible-haproxy_server stable/ocata: Added Haproxy global tunables https://review.openstack.org/446755 | 19:14 |
| openstackgerrit | Merged openstack/openstack-ansible-haproxy_server stable/newton: Added Haproxy global tunables https://review.openstack.org/446754 | 19:14 |
| agrebennikov | oh, wait... | 19:15 |
| agrebennikov | why?... | 19:15 |
| agrebennikov | odyssey4me, | 19:15 |
| agrebennikov | I was about to add the change :) | 19:15 |
| agrebennikov | sp is it going to be a separate commit now? | 19:17 |
| *** aludwar has quit IRC | 19:18 | |
| *** aludwar has joined #openstack-ansible | 19:19 | |
| *** Jack_Iv has quit IRC | 19:19 | |
| odyssey4me | agrebennikov this should be an edit to the release note in master | 19:19 |
| odyssey4me | it'll be a seperate commit | 19:20 |
| agrebennikov | aahh | 19:20 |
| odyssey4me | the backports are fine as-is as nothing has been removed, only added | 19:20 |
| odyssey4me | so in the master patch you've made the feature note which is fine as-is | 19:20 |
| odyssey4me | add another portion to the same release note with the heading 'upgrade: ' | 19:21 |
| agrebennikov | yep | 19:21 |
| odyssey4me | with text as described above | 19:21 |
| *** Andrew_jedi has quit IRC | 19:22 | |
| *** Jack_Iv has joined #openstack-ansible | 19:23 | |
| openstackgerrit | Andrey Grebennikov proposed openstack/openstack-ansible-haproxy_server master: Mentioned haproxy_bufsize option removal https://review.openstack.org/447104 | 19:24 |
| agrebennikov | odyssey4me, like this? https://review.openstack.org/447104 | 19:25 |
| odyssey4me | a few tweaks - can I make them for you? | 19:28 |
| agrebennikov | sure, thanks you very much! | 19:29 |
| openstackgerrit | Jesse Pretorius (odyssey4me) proposed openstack/openstack-ansible-haproxy_server master: Mentioned haproxy_bufsize option removal https://review.openstack.org/447104 | 19:30 |
| odyssey4me | there we go | 19:30 |
| odyssey4me | now when we release pike, anyone who previously tuned with that setting will know to change it | 19:31 |
| agrebennikov | exactly | 19:31 |
| *** Andrew_jedi has joined #openstack-ansible | 19:32 | |
| agrebennikov | sorry for my bad language (not native) :P | 19:32 |
| agrebennikov | cloudnull, can you please vote here? https://review.openstack.org/447104 | 19:33 |
| *** rboyapat has joined #openstack-ansible | 19:35 | |
| rboyapat | core members: can you please review https://review.openstack.org/#/c/446838/ | 19:35 |
| *** xinli has quit IRC | 19:35 | |
| *** manheim has joined #openstack-ansible | 19:41 | |
| openstackgerrit | Merged openstack/openstack-ansible-haproxy_server master: Mentioned haproxy_bufsize option removal https://review.openstack.org/447104 | 19:42 |
| *** shardy has quit IRC | 19:44 | |
| *** galstrom is now known as galstrom_zzz | 20:13 | |
| openstackgerrit | Kyle L. Henderson proposed openstack/openstack-ansible stable/newton: Provide example of using veth pairs for br-vlan https://review.openstack.org/445556 | 20:17 |
| *** foutatoro has quit IRC | 20:24 | |
| *** yolanda has quit IRC | 20:26 | |
| *** manheim has quit IRC | 20:26 | |
| *** yolanda has joined #openstack-ansible | 20:26 | |
| *** xinli has joined #openstack-ansible | 20:27 | |
| jrosser_ | odyssey4me: i see earlier you talked about removing ssh from the containers | 20:28 |
| *** cpuga has quit IRC | 20:29 | |
| *** cpuga has joined #openstack-ansible | 20:29 | |
| jrosser_ | is that in progress somewhere, as it would make stuff significantly easier in our situation | 20:29 |
| *** cpuga has quit IRC | 20:30 | |
| odyssey4me | jrosser_ it's not really a planned activity, primarily because no-one has really raised it as a priority | 20:30 |
| *** cpuga has joined #openstack-ansible | 20:30 | |
| *** retreved has quit IRC | 20:31 | |
| odyssey4me | we did a significant amount of work in Newton to make it possible, but the rest will require someone taking the time to figure out what breaks when openssh is removed from the containers and then it'll take a bit of collective thought to figure out what to do about those things | 20:31 |
| odyssey4me | if it's of interest to you, then perhaps you have resources to commit to doing the initial investigation? | 20:32 |
| odyssey4me | we could probably collaborate on figuring this stuff out on an etherpad, then work up a spec | 20:32 |
| *** crushil has quit IRC | 20:32 | |
| DimGR | i had major issues too with ssh into containers from not local host | 20:33 |
| odyssey4me | but most of us are already tied up with other activities, so we'll definitely need more hands | 20:33 |
| *** smatzek has quit IRC | 20:34 | |
| *** Jack_Iv has quit IRC | 20:34 | |
| *** cpuga has quit IRC | 20:35 | |
| *** chris_hultin is now known as chris_hultin|AWA | 20:36 | |
| *** cjloader has quit IRC | 20:40 | |
| *** crushil has joined #openstack-ansible | 20:42 | |
| *** Andrew_jedi has quit IRC | 20:44 | |
| jrosser_ | odyssey4me: can i ask about the ssh stuff then, currently the deployment host never ssh to the containers? | 20:46 |
| *** Jack_Iv has joined #openstack-ansible | 20:47 | |
| jrosser_ | but the containers need to ssh between themselves to distribute certs etc? | 20:47 |
| odyssey4me | currently the deployment host connects through the hosts to get to the containers for almost everything | 20:50 |
| odyssey4me | we have a connection plugin which makes Ansible connect to the host, then lxc-attach into the container to execute things | 20:51 |
| odyssey4me | therefore it does not need openssh running in the container | 20:51 |
| odyssey4me | however, we have delegated tasks - for example the os-glance-install.yml playbook targets the glance containers, but delegates the database creation to the galera containers and the rabbitmq vhost creation to the rabbitmq containers | 20:52 |
| odyssey4me | when that happens, it requires ssh connectivity between the containers | 20:52 |
| jrosser_ | ok | 20:52 |
| odyssey4me | to prevent that, we would have to change the pattern of how we do these things | 20:52 |
| jrosser_ | currently we have a service network with provisions all the bare metal, and has the deploment host on it | 20:53 |
| jrosser_ | this is not the mgmt network though | 20:53 |
| jrosser_ | and so we get into difficulty | 20:53 |
| *** Andrew_jedi has joined #openstack-ansible | 20:54 | |
| jrosser_ | but knowing that the service net does not need to ssh to the containers simplifies things | 20:54 |
| jrosser_ | it only has to go to all the hosts | 20:55 |
| odyssey4me | yeah | 20:55 |
| odyssey4me | I'm not sure what would fall out if you do this, but you could also add a different bridge and apply these settings on it instead of on br-mgmt: https://github.com/openstack/openstack-ansible/blob/master/etc/openstack_deploy/openstack_user_config.yml.prod.example#L34-L35 | 20:56 |
| jrosser_ | yes we could | 20:56 |
| jrosser_ | but the point of the service net is it is more trusted than the mgmt net | 20:56 |
| odyssey4me | there's a fair chance that we've made a lot of assumptions about br-mgmt being the network ansible connects to, but that is stuff we should actually clean up | 20:56 |
| jrosser_ | and very deliberatley isolated from it | 20:56 |
| odyssey4me | yep, fair point | 20:57 |
| jrosser_ | i could never get the deployment host being on mgmt past a security review | 20:57 |
| odyssey4me | br-mgmt is meant to be an isolated network with only outgoing internet access anyway | 20:58 |
| odyssey4me | it's probably a badly named network | 20:58 |
| jrosser_ | indeed :) | 20:58 |
| jrosser_ | im also trying not to have mgmt having outgoing internet access | 20:59 |
| odyssey4me | with or without a proxy | 20:59 |
| jrosser_ | proxy is ok | 20:59 |
| jrosser_ | nat to internet is not | 21:00 |
| odyssey4me | ok, that should work already | 21:00 |
| jrosser_ | yes | 21:00 |
| odyssey4me | it's a bit of a PITA to get right in the beginning, but it works eventually | 21:00 |
| jrosser_ | we are just crawling from multinode aio to something like that | 21:00 |
| odyssey4me | I'm doing a bunch of work to try and make a fully air-gapped deployment work. | 21:01 |
| jrosser_ | but as you say it is a lot of work to get the environment right | 21:01 |
| jrosser_ | oh ++ on that | 21:01 |
| jrosser_ | happy to test things | 21:01 |
| odyssey4me | yeah, once we have the whole story worked out and broader testing started in our group we'll start to figure out how to upstream the tooling | 21:01 |
| odyssey4me | a lot of it right now is a bit of a hack job | 21:02 |
| jrosser_ | ideally my service net provides all proxies and packeges etc | 21:02 |
| *** xinli has quit IRC | 21:02 | |
| jrosser_ | and mgmt becomes purely api chatter | 21:03 |
| *** crushil has quit IRC | 21:03 | |
| jrosser_ | but at the same time i dont want the service net in the containers | 21:03 |
| odyssey4me | but the basic stuff we're doing it preparing all packages, repo data, etc in a pipeline - then tooling the deployment to stage either offline or online, then execute the normal build using the staged data | 21:03 |
| jrosser_ | so thats a bit chicken/egg atm | 21:03 |
| *** rboyapat has quit IRC | 21:04 | |
| odyssey4me | yeah, it'd be nice to figure that out - I expect it can be done with what's already there - perhaps with a few patches to ensure we're looking the right things up the right way | 21:04 |
| *** rboyapat has joined #openstack-ansible | 21:04 | |
| *** McMurlock1 has quit IRC | 21:06 | |
| *** rboyapat has quit IRC | 21:07 | |
| jrosser_ | i also need to think about the implications of things like bare metal network nodes where the service net and provider nets come together with little isolation | 21:08 |
| jrosser_ | thats feeling a little less separated than regualar infra nodes currently | 21:08 |
| *** foutatoro has joined #openstack-ansible | 21:19 | |
| *** cjloader has joined #openstack-ansible | 21:24 | |
| *** cjloader has quit IRC | 21:28 | |
| odyssey4me | yeah, unfortunately ironic doesn't really separate things properly just yet as far as I know - it puts all projects on the same provisioning network, so you'd have to do a ton of ACL's to secure it | 21:29 |
| *** Andrew_jedi has quit IRC | 21:32 | |
| *** kstev has quit IRC | 21:32 | |
| *** gouthamr has quit IRC | 21:33 | |
| *** Jack_Iv has quit IRC | 21:34 | |
| *** Jack_Iv has joined #openstack-ansible | 21:35 | |
| *** Jeffrey4l has quit IRC | 21:35 | |
| *** schwicht has quit IRC | 21:36 | |
| *** Jeffrey4l has joined #openstack-ansible | 21:36 | |
| *** Jack_Iv has quit IRC | 21:39 | |
| *** esberglu has quit IRC | 21:41 | |
| *** esberglu has joined #openstack-ansible | 21:41 | |
| *** esberglu has quit IRC | 21:46 | |
| *** karimb has quit IRC | 21:51 | |
| *** gouthamr has joined #openstack-ansible | 21:57 | |
| openstackgerrit | German Eichberger proposed openstack/openstack-ansible-os_octavia master: [WIP] Adds iptables rules to protect octavia server container https://review.openstack.org/447151 | 22:03 |
| dankolbrs | Hi all, I submitted https://bugs.launchpad.net/openstack-ansible/+bug/1673889 . Feel free to yell at me here or in there if I missed anything or left anything out | 22:05 |
| openstack | Launchpad bug 1673889 in openstack-ansible "Nova services do not restart on N->O upgrade" [Undecided,New] | 22:05 |
| cmart | odyssey4me and agrebennikov, I solved the libvirt key injection issue. in Ubuntu the kernel is non-readable to non-root users by default, so nova could not read the kernel and launch the supermin! | 22:21 |
| *** schwicht has joined #openstack-ansible | 22:21 | |
| odyssey4me | cmart interesting - so how did you solve it? does it need a rootwrap edit? | 22:21 |
| cmart | well just now I solved it with `sudo chmod 0644 /boot/vmlinuz*` | 22:22 |
| odyssey4me | if so, then it's actually a nova bug which you can patch :) | 22:22 |
| odyssey4me | hmm, I expect that nova may be interested in it - if you could put together a bug report to describe the error output, the config bits and the packages/platform then we could perhaps engage with some of the nova crew to see what to do about it | 22:23 |
| odyssey4me | johnthetubaguy may have some insight | 22:23 |
| *** foutatoro has quit IRC | 22:25 | |
| cmart | ok. it is an Ubuntu-specific bug. something needs to make sure that the user running libguestfs (in our case nova) can read the compressed kernel files in /boot. it sounds like there have been years of handwringing over this between libguestfs, Ubuntu, and various other projects | 22:25 |
| odyssey4me | cmart sure, but that's exactly the purpose of the rootwrap files | 22:25 |
| odyssey4me | they're effectively files which implement sudoer capabilities | 22:26 |
| cmart | is rootwrap a nova-ism or an OSA-ism? | 22:26 |
| odyssey4me | https://github.com/openstack/nova/blob/master/etc/nova/rootwrap.d/compute.filters | 22:26 |
| cmart | aha | 22:26 |
| odyssey4me | rootwrap itself is a python wrapper which is implemented for priveleged commands - it reads that file to figure out what it's allowed to do, or not | 22:27 |
| odyssey4me | but it's basically a sudoers file | 22:27 |
| cmart | ok. and if the config for "what can run as root" is maintained in nova, i'll write up the bug against nova. should be a simple fix. Thanks odyssey4me and agrebennikov for helping me triage! | 22:29 |
| odyssey4me | cmart note this though: https://github.com/openstack/nova/blob/424972e2f03c42b76d47775b016d3ec2d001632f/nova/conf/libvirt.py#L147-L153 | 22:29 |
| cmart | ya - what about? | 22:30 |
| odyssey4me | actually, more importantly https://github.com/openstack/nova/blob/424972e2f03c42b76d47775b016d3ec2d001632f/nova/conf/libvirt.py#L169-L181 | 22:30 |
| odyssey4me | libguestfs appears to be broken, so there's that | 22:30 |
| odyssey4me | but if libguestfs is not there, nbd will be used | 22:30 |
| odyssey4me | the long term fix is to get the libguestfs fix into nova | 22:31 |
| odyssey4me | the short term fix may be to make sure libguestfs is not there | 22:31 |
| agrebennikov | cmart, nice stuff :) | 22:31 |
| odyssey4me | if OSA puts it there, we can remove that quickly | 22:31 |
| odyssey4me | any patches into nova will take time, and are not likely to be backported | 22:31 |
| odyssey4me | of course we'd like you to validate that removing it helps :) | 22:32 |
| cmart | unfortunately nbd as a fallback doesn't help you if you set inject_partition to -1, meaning libguestfs must be available to find the correct partition to mount https://github.com/openstack/nova/blob/424972e2f03c42b76d47775b016d3ec2d001632f/nova/conf/libvirt.py#L185-L198 | 22:32 |
| agrebennikov | odyssey4me, "patches into nova will take time" should be said like "patches into nova will take forever" | 22:32 |
| agrebennikov | :P | 22:32 |
| odyssey4me | :) I'm trying to be polite agrebennikov - the nova crew have the hardest job. | 22:33 |
| cmart | (and that's how I'm doing it!) | 22:33 |
| cmart | ok. thanks guys. I'll get this written up later tonight. gotta scoot and get my taxes done. | 22:33 |
| odyssey4me | cmart if anything, you may get advise for alternative ways of achieving the same thing | 22:34 |
| agrebennikov | cmart, have fun with taxes ;) | 22:34 |
| odyssey4me | happy ot have helped - have a great w/end! | 22:34 |
| agrebennikov | odyssey4me, before you leave - can you please look at http://paste.openstack.org/show/603198/ and tell me if it can be accepted as an extention to nova.conf? | 22:35 |
| agrebennikov | or there is easier way for doing it | 22:36 |
| agrebennikov | it is regarding https://bugs.launchpad.net/openstack-ansible/+bug/1673570 | 22:36 |
| openstack | Launchpad bug 1673570 in openstack-ansible "With more than one ceph cluster as the backend nova has secret_uuid hardcoded" [Undecided,New] | 22:36 |
| agrebennikov | (only in case you are familiar with that part) | 22:36 |
| agrebennikov | if not - lets go home | 22:37 |
| *** vnogin has quit IRC | 22:38 | |
| odyssey4me | argh, the whole usage of ceph needs a rewrite | 22:40 |
| odyssey4me | unfortunately I don't understand that code, and certainly won't right now - I'm quite tired | 22:40 |
| odyssey4me | but now that we have everything integrated we can rewrite all that nonsense which hasn't changed much since kilo | 22:41 |
| odyssey4me | it's time we do that | 22:41 |
| odyssey4me | to fairly review this I'd have to try and work out the spider web | 22:41 |
| odyssey4me | next week we can try and see if logan- or mattt are available to help puzzle it out - they understand it better than most | 22:42 |
| logan- | i'll read thru it tonight | 22:42 |
| odyssey4me | logan- actually uses it in production, so that helps | 22:42 |
| odyssey4me | lol, a logan- lurks in the bushes | 22:42 |
| logan- | odyssey4me ;) someone said ceph | 22:42 |
| odyssey4me | hahaha | 22:43 |
| odyssey4me | how's your upgrade to newton going? | 22:43 |
| odyssey4me | sorry, the upgrade from trusty to xenial | 22:43 |
| logan- | 100% done now :) | 22:43 |
| logan- | no more trusty anywhere | 22:43 |
| odyssey4me | oh awesome :) good news, well worth a celebration | 22:44 |
| agrebennikov | all right, logan- can I bug you on monday morning then? | 22:44 |
| odyssey4me | time to look forward again | 22:44 |
| odyssey4me | well - probably time to take a break, actually | 22:44 |
| logan- | no kidding! great to have it behind. really cool work cloudnull did this week with the repo stuff. I was just reading some of that today | 22:44 |
| logan- | agrebennikov: i've got a question for you, 1 sec let me find this older review | 22:45 |
| agrebennikov | absolutely | 22:45 |
| agrebennikov | topic? | 22:45 |
| agrebennikov | I have all mine opened | 22:45 |
| logan- | https://review.openstack.org/#/c/409353/ is your bug/patch similar to this one? | 22:46 |
| agrebennikov | soooooo not actually :) | 22:46 |
| odyssey4me | hmm, yeah - it was michaelgugino who was working on ceph multi back-end things | 22:46 |
| agrebennikov | in fact it ia all messed up, and I'd say on the nova side | 22:46 |
| odyssey4me | unfortunately we haven't seen him for a while | 22:46 |
| logan- | i haven't really read thru your bug yet but i was just curious if its the same/similar thing | 22:46 |
| agrebennikov | the main idea - with multi-ceph AND ephemerals in ceph nova can't work | 22:47 |
| odyssey4me | yeah, I think it's time to w/end y'all | 22:47 |
| agrebennikov | and my past above should cover that | 22:47 |
| agrebennikov | yeah, that makes perfect sense | 22:47 |
| logan- | ttyl odyssey4me | 22:47 |
| logan- | agrebennikov: why doesn't it work? missing libvirt secrets? | 22:48 |
| agrebennikov | ok, long story short - with volumes in ceph secret uuid and ceph user can be delivered via rpc from cinder | 22:48 |
| logan- | nevermind.. i'll read the bug and patch(es) first and bug you with questions afterwards :P | 22:49 |
| agrebennikov | with ephemerals in ceph these params Must be hardcoded in nova.conf on the compute | 22:49 |
| logan- | right | 22:49 |
| agrebennikov | I have to submit a couple of new params into nova though | 22:49 |
| agrebennikov | but it is a separate story | 22:49 |
| agrebennikov | let's chat on monday | 22:49 |
| logan- | ok sounds good | 22:50 |
| logan- | have a good weekend | 22:50 |
| agrebennikov | nice. cu then! have a great green-beer evening! | 22:50 |
| logan- | yes indeed | 22:50 |
| agrebennikov | (if you are here in the us) | 22:50 |
| *** schwicht has quit IRC | 22:51 | |
| logan- | dallas | 22:51 |
| logan- | :) | 22:51 |
| *** agrebennikov has quit IRC | 22:55 | |
| openstackgerrit | Merged openstack/openstack-ansible-rsyslog_client master: Fix remote logging template https://review.openstack.org/446838 | 22:56 |
| *** jbadiapa has quit IRC | 23:01 | |
| *** vnogin has joined #openstack-ansible | 23:02 | |
| *** jamesdenton has joined #openstack-ansible | 23:12 | |
| *** jamesden_ has joined #openstack-ansible | 23:16 | |
| *** schwicht has joined #openstack-ansible | 23:17 | |
| *** jamesdenton has quit IRC | 23:19 | |
| *** markvoelker has quit IRC | 23:26 | |
| *** DimGR has quit IRC | 23:43 | |
| *** vnogin has quit IRC | 23:47 | |
| *** marst has joined #openstack-ansible | 23:50 | |
| *** crushil has joined #openstack-ansible | 23:52 | |
| *** acormier has joined #openstack-ansible | 23:53 | |
| *** acormier has quit IRC | 23:55 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!