| *** cloudnautique has quit IRC | 00:07 | |
| *** suro-patz1 has quit IRC | 00:10 | |
| *** suro-patz has joined #kolla | 00:11 | |
| *** thumpba has quit IRC | 00:11 | |
| *** thumpba has joined #kolla | 00:12 | |
| *** suro-patz has quit IRC | 00:12 | |
| *** suro-patz has joined #kolla | 00:12 | |
| *** thumpba has quit IRC | 00:16 | |
| sdake | samyaple why did you chose source kolla_extend_start rather then sudo -E kolla_extend_start? | 00:18 |
|---|---|---|
| *** dims has joined #kolla | 00:20 | |
| *** thumpba has joined #kolla | 00:22 | |
| *** vilobhmm has joined #kolla | 00:31 | |
| *** dwalsh has joined #kolla | 00:37 | |
| *** rhallisey has quit IRC | 00:52 | |
| *** jasonsb_ has quit IRC | 00:55 | |
| *** jasonsb has joined #kolla | 00:55 | |
| *** jasonsb has quit IRC | 01:00 | |
| *** vilobhmm1 has joined #kolla | 01:01 | |
| *** weiyu has joined #kolla | 01:02 | |
| *** vilobhmm has quit IRC | 01:03 | |
| *** thumpba_ has joined #kolla | 01:05 | |
| *** thumpba has quit IRC | 01:05 | |
| *** thumpba has joined #kolla | 01:12 | |
| *** thumpba_ has quit IRC | 01:13 | |
| *** thumpba has quit IRC | 01:13 | |
| *** thumpba has joined #kolla | 01:13 | |
| *** weiyu has quit IRC | 01:16 | |
| *** signed8bit_ZZZzz is now known as signed8bit | 01:19 | |
| *** signed8bit has quit IRC | 01:19 | |
| *** weiyu_ has joined #kolla | 01:20 | |
| *** weiyu_ has quit IRC | 01:30 | |
| *** weiyu_ has joined #kolla | 01:31 | |
| *** dims has quit IRC | 01:34 | |
| *** suro-patz1 has joined #kolla | 01:35 | |
| *** suro-patz has quit IRC | 01:36 | |
| *** achanda has joined #kolla | 01:47 | |
| *** weiyu_ has quit IRC | 01:51 | |
| *** weiyu has joined #kolla | 02:00 | |
| *** thumpba_ has joined #kolla | 02:09 | |
| *** thumpba has quit IRC | 02:09 | |
| *** thumpba_ has quit IRC | 02:09 | |
| *** thumpba has joined #kolla | 02:10 | |
| *** suro-patz1 has quit IRC | 02:11 | |
| *** thumpba has quit IRC | 02:13 | |
| *** dims has joined #kolla | 02:22 | |
| *** stvnoyes has quit IRC | 02:23 | |
| *** stvnoyes has joined #kolla | 02:23 | |
| *** ssurana has quit IRC | 02:25 | |
| *** sacharya has joined #kolla | 02:27 | |
| *** sacharya has quit IRC | 02:32 | |
| *** weiyu has quit IRC | 02:46 | |
| *** kjelly_ is now known as kjelly | 03:02 | |
| *** dims has quit IRC | 03:07 | |
| *** cloudnautique has joined #kolla | 03:08 | |
| *** dwalsh has quit IRC | 03:11 | |
| *** cloudnautique has quit IRC | 03:13 | |
| *** vilobhmm1 has quit IRC | 03:20 | |
| *** tummy has quit IRC | 03:44 | |
| *** achanda has quit IRC | 03:48 | |
| *** achanda has joined #kolla | 03:49 | |
| *** cemmason has quit IRC | 03:51 | |
| *** cemmason has joined #kolla | 03:51 | |
| *** cemmason has quit IRC | 03:52 | |
| *** cemmason has joined #kolla | 03:52 | |
| *** vbel has quit IRC | 04:01 | |
| *** vbel has joined #kolla | 04:02 | |
| *** tfukushima has joined #kolla | 04:02 | |
| *** tfukushima has quit IRC | 04:05 | |
| *** sdake has quit IRC | 04:14 | |
| *** alisonh has quit IRC | 04:28 | |
| *** weiyu_ has joined #kolla | 04:29 | |
| *** alisonh has joined #kolla | 04:54 | |
| *** achanda has quit IRC | 04:54 | |
| *** sacharya has joined #kolla | 04:55 | |
| *** jasonsb has joined #kolla | 04:57 | |
| *** ubuntu1 has joined #kolla | 05:09 | |
| *** kragniz_ has joined #kolla | 05:09 | |
| *** another_larsks has joined #kolla | 05:13 | |
| *** thomasem_ has joined #kolla | 05:14 | |
| *** Windir has quit IRC | 05:14 | |
| *** kragniz has quit IRC | 05:14 | |
| *** akscram has quit IRC | 05:14 | |
| *** larsks has quit IRC | 05:14 | |
| *** thomasem has quit IRC | 05:14 | |
| *** anteaya has quit IRC | 05:14 | |
| *** akscram has joined #kolla | 05:15 | |
| *** Windir_ has joined #kolla | 05:15 | |
| *** kjelly has quit IRC | 05:26 | |
| *** weiyu_ has quit IRC | 05:30 | |
| *** Windir_ is now known as Windir | 05:31 | |
| *** harmw has joined #kolla | 05:49 | |
| *** harmw_ has quit IRC | 05:51 | |
| *** rbradfor has quit IRC | 05:51 | |
| *** odyssey4me has quit IRC | 05:51 | |
| *** ArchiFleKs has quit IRC | 05:51 | |
| *** odyssey4me has joined #kolla | 05:51 | |
| *** kragniz_ is now known as kragniz | 05:51 | |
| *** ArchiFleKs has joined #kolla | 05:52 | |
| *** rbradfor has joined #kolla | 05:53 | |
| *** achanda has joined #kolla | 06:02 | |
| *** suro-patz has joined #kolla | 06:05 | |
| *** suro-patz has joined #kolla | 06:06 | |
| *** achanda has quit IRC | 06:07 | |
| *** pbourke has quit IRC | 06:17 | |
| *** sacharya_ has joined #kolla | 06:17 | |
| *** pbourke has joined #kolla | 06:17 | |
| *** bradjones has quit IRC | 06:17 | |
| *** bradjones has joined #kolla | 06:19 | |
| *** bradjones has quit IRC | 06:19 | |
| *** bradjones has joined #kolla | 06:19 | |
| *** sacharya has quit IRC | 06:20 | |
| *** weiyu_ has joined #kolla | 06:36 | |
| *** sacharya has joined #kolla | 06:36 | |
| *** achanda has joined #kolla | 06:39 | |
| *** jasonsb has quit IRC | 06:39 | |
| *** sacharya_ has quit IRC | 06:39 | |
| *** suro-patz has quit IRC | 06:45 | |
| *** achanda has quit IRC | 06:49 | |
| *** jasonsb has joined #kolla | 06:53 | |
| *** weiyu_ has quit IRC | 06:57 | |
| *** weiyu_ has joined #kolla | 06:58 | |
| *** weiyu_ has quit IRC | 07:01 | |
| *** SamYaple_ has joined #kolla | 07:03 | |
| *** clayton has quit IRC | 07:03 | |
| *** SamYaple has quit IRC | 07:03 | |
| *** clayton has joined #kolla | 07:04 | |
| *** jasonsb has quit IRC | 07:11 | |
| *** weiyu_ has joined #kolla | 07:19 | |
| *** weiyu_ has quit IRC | 07:23 | |
| *** weiyu_ has joined #kolla | 07:25 | |
| *** weiyu_ has quit IRC | 07:27 | |
| *** weiyu has joined #kolla | 07:35 | |
| *** weiyu has quit IRC | 07:50 | |
| *** weiyu_ has joined #kolla | 07:54 | |
| *** stvnoyes has quit IRC | 07:57 | |
| *** stvnoyes has joined #kolla | 07:59 | |
| *** sdake has joined #kolla | 08:08 | |
| *** sdake_ has joined #kolla | 08:12 | |
| *** sdake has quit IRC | 08:13 | |
| *** sdake_ has quit IRC | 08:13 | |
| *** weiyu_ has quit IRC | 08:21 | |
| *** weiyu has joined #kolla | 08:44 | |
| *** achanda has joined #kolla | 08:54 | |
| *** achanda has quit IRC | 08:56 | |
| *** dwalsh has joined #kolla | 09:29 | |
| *** weiyu has quit IRC | 09:35 | |
| *** weiyu has joined #kolla | 09:36 | |
| *** weiyu has quit IRC | 09:37 | |
| *** sacharya has quit IRC | 09:40 | |
| *** weiyu has joined #kolla | 09:40 | |
| *** weiyu has quit IRC | 09:44 | |
| *** exploreshaifali has joined #kolla | 09:46 | |
| *** weiyu_ has joined #kolla | 09:47 | |
| *** weiyu_ has quit IRC | 09:50 | |
| *** ryansb_ has joined #kolla | 10:18 | |
| *** alisonh1 has joined #kolla | 10:18 | |
| *** ryansb_ has quit IRC | 10:18 | |
| *** ryansb_ has joined #kolla | 10:18 | |
| *** SamYaple has joined #kolla | 10:19 | |
| *** barra204 has joined #kolla | 10:20 | |
| *** bigjools_ has joined #kolla | 10:20 | |
| *** sdake has joined #kolla | 10:20 | |
| *** bigjools_ has quit IRC | 10:20 | |
| *** akscram has quit IRC | 10:20 | |
| *** alisonh has quit IRC | 10:20 | |
| *** SamYaple_ has quit IRC | 10:20 | |
| *** another_larsks has quit IRC | 10:20 | |
| *** whydidyoustealmy has quit IRC | 10:20 | |
| *** openstackgerrit has quit IRC | 10:20 | |
| *** bigjools has quit IRC | 10:20 | |
| *** ryansb has quit IRC | 10:20 | |
| *** ryansb_ is now known as ryansb | 10:20 | |
| *** bigjools has joined #kolla | 10:20 | |
| *** bigjools has quit IRC | 10:20 | |
| *** bigjools has joined #kolla | 10:20 | |
| *** akscram has joined #kolla | 10:21 | |
| *** openstackgerrit has joined #kolla | 10:22 | |
| *** larsks has joined #kolla | 10:24 | |
| *** weiyu_ has joined #kolla | 10:27 | |
| *** sdake has quit IRC | 10:33 | |
| *** sdake has joined #kolla | 10:36 | |
| *** weiyu_ has quit IRC | 10:41 | |
| *** weiyu_ has joined #kolla | 10:41 | |
| *** sdake has quit IRC | 10:54 | |
| *** alisonh1 has quit IRC | 10:59 | |
| *** weiyu_ has quit IRC | 11:15 | |
| *** alisonh has joined #kolla | 11:19 | |
| *** kjelly has joined #kolla | 11:38 | |
| *** dwalsh has quit IRC | 11:47 | |
| *** jmccarthy has quit IRC | 12:38 | |
| *** jmccarthy has joined #kolla | 12:39 | |
| *** kjelly has quit IRC | 13:10 | |
| *** slagle has quit IRC | 13:19 | |
| *** slagle has joined #kolla | 13:26 | |
| *** dims has joined #kolla | 13:37 | |
| *** achanda has joined #kolla | 13:39 | |
| *** exploreshaifali has quit IRC | 13:48 | |
| *** achanda has quit IRC | 13:50 | |
| *** achanda has joined #kolla | 13:51 | |
| *** achanda has quit IRC | 13:54 | |
| *** achanda has joined #kolla | 13:54 | |
| *** slagle has quit IRC | 13:55 | |
| *** achanda has quit IRC | 13:59 | |
| *** cemmason1 has joined #kolla | 14:30 | |
| *** cemmason has quit IRC | 14:33 | |
| openstackgerrit | Manoj Kumar Jain proposed openstack/kolla: Improvement in deploy-all-in-on-node.rst Modified doc for deploying kolla on all distro platform Add steps to find the correct configuration parameters Add steps to verify if build and deployment is successful https://review.openstack.org/236667 | 14:47 |
| *** slagle has joined #kolla | 15:12 | |
| *** docaedo is now known as argentina | 15:14 | |
| openstackgerrit | Britt Houser proposed openstack/kolla: drop root on memcached https://review.openstack.org/245330 | 15:14 |
| *** argentina is now known as docaedo | 15:14 | |
| openstackgerrit | Britt Houser proposed openstack/kolla: drop root on memcached https://review.openstack.org/245330 | 15:16 |
| *** slagle has quit IRC | 15:33 | |
| *** suro-patz has joined #kolla | 15:56 | |
| *** dims has quit IRC | 16:01 | |
| *** sdake has joined #kolla | 16:19 | |
| *** sdake has quit IRC | 16:24 | |
| *** sdake has joined #kolla | 16:28 | |
| britthouser | SamYaple - its definitely the right user (memcached). But I wonder if the {{user }} replacement isn't right? | 16:36 |
| britthouser | oh I think I found it | 16:38 |
| britthouser | didn't single quote the user | 16:38 |
| openstackgerrit | Britt Houser proposed openstack/kolla: drop root on memcached https://review.openstack.org/245330 | 16:38 |
| *** sdake has quit IRC | 16:40 | |
| *** sdake has joined #kolla | 16:41 | |
| sdake | morning | 16:50 |
| sdake | britthouser i looked at the gatea | 16:50 |
| sdake | and usesrmod was throwing an error | 16:50 |
| sdake | perhaps that %set isn't doing what you think it is doing | 16:50 |
| sdake | or {{ user }} does do what you think it does | 16:50 |
| sdake | so see the last gat run, click "toggle ci" | 16:50 |
| sdake | britthouser http://logs.openstack.org/30/245330/4/check/gate-kolla-dsvm-build-centos-binary/a09cd2a/console.html#_2015-11-14_15_22_08_410 | 16:51 |
| britthouser | Yeah I hadn't single quoted the user when I did the set | 16:52 |
| britthouser | I'm pretty sure that is the issue...waiting for the gate on patch 5 | 16:52 |
| sdake | so britt | 16:52 |
| sdake | on a scale of 1 to 10 | 16:52 |
| sdake | howo much ass does our gaating kick | 16:52 |
| sdake | 10 being ass beating | 16:52 |
| sdake | 1 = opppoite ;) | 16:53 |
| britthouser | That is a tough one. It definitely kicked mine, but I don't think mine qualifies as a 10...more like a three. But it probably could kick a 10 as well. =) | 16:54 |
| sdake | what i mean is how good do you thiink it is compared to - say mercs gating | 16:54 |
| britthouser | Oh...way better. But that's something on the plate for this sprint. | 16:55 |
| sdake | it is not achieavable in one sprint | 16:55 |
| britthouser | we're setting up multi-node deployment on metal. | 16:55 |
| sdake | sam and I have spent hundreds of hours each on the gating | 16:55 |
| sdake | thats good, we will need an internal gate | 16:56 |
| britthouser | I'm sure the deploy gating a better, but I'd have to give our build gating the upper hand. | 16:56 |
| britthouser | but we don't have the pemutations | 16:56 |
| sdake | could you explain why | 16:56 |
| sdake | (just so we can fil lgaps) | 16:56 |
| britthouser | We have tooling to only build containers that are changed. We can trigger container build not just on dockerfiles, but RPM spec as well as source trees. And we serverspec each container at the end. | 16:57 |
| sdake | how do you mean sourcee trees | 16:58 |
| sdake | i don tthink there is a whole lot of value in serverspecing a built container | 16:59 |
| *** suro-patz has quit IRC | 16:59 | |
| britthouser | someone modifies neutron source, triggers RPM build, then we see which containers need that package, build those containers, etc. | 16:59 |
| sdake | building only changed containers is suboptimal - need to test everything as one unit | 16:59 |
| sdake | justi mo ;) | 16:59 |
| sdake | imo ;) | 16:59 |
| sdake | why not just rebuild it all? | 17:00 |
| britthouser | time. | 17:00 |
| sdake | takes us 20-25 minutes for the gate to run | 17:00 |
| sdake | all jobs run in parallell | 17:00 |
| sdake | sometimes if openstack is busy that is not the case | 17:00 |
| sdake | but generally tis true | 17:01 |
| britthouser | Yeah maybe our CI cloud is underpowered? but it was painful to build world. | 17:01 |
| sdake | that will be easy enough to integrate with (specific buidls) - our build script supports regex building | 17:01 |
| britthouser | So we can build a subset, b/c we push registry, and then tag the container we built, plus all teh containers from the same source sha together. | 17:02 |
| britthouser | so we always have a full set, even with partial build. | 17:02 |
| britthouser | serverspec on containers has actually helped. | 17:03 |
| sdake | in which way | 17:03 |
| sdake | we never have missing content from our containers | 17:03 |
| britthouser | package naming changes sometimes. | 17:03 |
| sdake | that is caught by th build script erroring | 17:03 |
| sdake | not trying to be onery, just dont see the value :) | 17:04 |
| SamYaple | britthouser: there is no such things as "only containers that are changed" unfortunately | 17:04 |
| sdake | chris said the new build was down under 10 hours? | 17:04 |
| SamYaple | the thing is you have to update your package lists | 17:04 |
| SamYaple | thats done one time (in the base) to ensure all packages are the same | 17:05 |
| SamYaple | otherwise different contaienrs might have different libraries | 17:05 |
| britthouser | Its caught bugs for us...but we didn't have a build-gate at the time. | 17:05 |
| SamYaple | they would _probably_ still work together | 17:05 |
| SamYaple | but the whole ideas to to ensure consistency | 17:05 |
| sdake | ya im not to oconcerned about different librarie for different containers | 17:05 |
| sdake | but it is better if eerything is concsistent | 17:06 |
| SamYaple | you arent concerned until it breaks | 17:07 |
| britthouser | I'm not quite following you guys there. | 17:07 |
| SamYaple | which it will and does | 17:07 |
| sdake | hey SamYaple do we have enough horsepower (ram) n the gate to start a vm via nova | 17:07 |
| SamYaple | nope | 17:07 |
| SamYaple | not with heat | 17:07 |
| SamYaple | its a killer | 17:07 |
| sdake | i wonder why | 17:07 |
| SamYaple | idk | 17:07 |
| britthouser | sdake - gate on patch 5 passes | 17:07 |
| SamYaple | it was eating all the rams | 17:07 |
| sdake | britt cool i'll ack | 17:08 |
| sdake | after i take another look | 17:08 |
| britthouser | thx | 17:08 |
| SamYaple | sdake: you ever goign to tag 1.0.1 | 17:08 |
| SamYaple | been waiting for a while here | 17:09 |
| SamYaple | you promised it like a week ago | 17:09 |
| SamYaple | 8 patches in the queue waiting for that tag... | 17:09 |
| sdake | well li guess the extra usermod is better then an extra layer | 17:09 |
| sdake | SamYaple i will tag this morning after i build and test ok? | 17:09 |
| SamYaple | sure | 17:09 |
| sdake | if you could be a darling and test ubuntu for me i'd superapprecaite it | 17:09 |
| sdake | just so we eknow everythign works | 17:10 |
| sdake | ok pulling and building for centos | 17:10 |
| sdake | build and test takes about an hour | 17:10 |
| sdake | remembering my super complex ssh key takes about 30 minutes | 17:10 |
| sdake | or gpg key that is | 17:10 |
| britthouser | do I need to backport the drop root? I can't remember is that was on teh list? | 17:11 |
| openstackgerrit | Merged openstack/kolla: drop root on memcached https://review.openstack.org/245330 | 17:11 |
| sdake | no | 17:11 |
| sdake | that is new work | 17:11 |
| britthouser | Ok cool. | 17:11 |
| sdake | nice job britt! | 17:12 |
| britthouser | Thanks!! | 17:12 |
| sdake | keep em coming | 17:12 |
| sdake | see, that wasn' t that hard was it | 17:12 |
| britthouser | Appreciate all the help. | 17:12 |
| sdake | tell your peers pleasse | 17:12 |
| britthouser | Well, I got alot of help | 17:12 |
| sdake | yes that is what we do, help each other | 17:12 |
| britthouser | zaqar was done by one of my pers | 17:12 |
| sdake | except sam and I, we mostly bicker :) | 17:12 |
| sdake | i think the zaqar container is fucked up | 17:12 |
| sdake | it is missing a usermod | 17:12 |
| britthouser | oops | 17:12 |
| sdake | i dont know why it got approved | 17:12 |
| SamYaple | i could agree with you sdake, but then we would both be wrong | 17:12 |
| sdake | sam lol | 17:13 |
| sdake | SamYaple was that a yes on testing ubuntu? | 17:13 |
| sdake | bulding stable/liberty now | 17:13 |
| SamYaple | sdake: zaqar got approved because it passed the gate | 17:13 |
| SamYaple | there is no functinoal for it, so it built but past that ehh | 17:13 |
| sdake | well i wish people would actually look at the patches | 17:13 |
| SamYaple | I looked at it | 17:14 |
| SamYaple | it looked fine | 17:14 |
| sdake | it is missinga the usermod | 17:14 |
| sdake | iirc | 17:14 |
| SamYaple | i get that, but still | 17:14 |
| SamYaple | i made sure the gate was happy | 17:14 |
| sdake | ya priority #1 is not breaking the gate | 17:14 |
| * sdake wtb voting gates | 17:14 | |
| SamYaple | the truth is that container wont work, if i remeber right it is missing some deps | 17:14 |
| SamYaple | so the whole container is busted anyway | 17:14 |
| sdake | ya we will tackle zaqar when we ansible it | 17:14 |
| sdake | i shoud lhave jusst left those off that we don't have anible for | 17:15 |
| sdake | but i guess some code in there is better then none | 17:15 |
| SamYaple | sdake: if you want to do more in the gate, start looking at how the services could consume less ram | 17:15 |
| sdake | i havent looked at the deploy code | 17:15 |
| sdake | are you deploying ceph and cinder? | 17:15 |
| sdake | we could problby run less proccesses for heat-engine and heat-api | 17:16 |
| sdake | i suspect it runs a whoole slew of them | 17:16 |
| openstackgerrit | Britt Houser proposed openstack/kolla: Add usermod to Zaqar container https://review.openstack.org/245436 | 17:16 |
| sdake | britt yay | 17:16 |
| britthouser | Monday I'd like to pick yoru brain about dev workflow on metal. | 17:17 |
| sdake | -1 on your review | 17:17 |
| sdake | pick it now | 17:17 |
| sdake | unless you hae somewhere to be | 17:17 |
| SamYaple | sdake: no ceph or cinder yet | 17:17 |
| SamYaple | no ram | 17:17 |
| sdake | we are getting ready to multi-node our deployment tests as well | 17:17 |
| openstackgerrit | Britt Houser proposed openstack/kolla: Add usermod to Zaqar container https://review.openstack.org/245436 | 17:17 |
| britthouser | sdake - gotta clean house for birthday party tommorow | 17:18 |
| britthouser | can we do monday? | 17:18 |
| sdake | britthouser enjoy | 17:18 |
| sdake | SamYaple when I run kolla it takes 8gb on bare metal | 17:18 |
| sdake | for a ocntrol node | 17:18 |
| sdake | its a shame more ram is a nonstarter | 17:18 |
| sdake | what about mor eswap? | 17:19 |
| sdake | we takled about that | 17:19 |
| SamYaple | im sure we can limit the ram | 17:19 |
| SamYaple | swap will make gate timeout | 17:19 |
| SamYaple | im not have 2 hour gates | 17:19 |
| sdake | the only thing i can think of is to reduce the number eof processes we run for the api processes etc | 17:19 |
| SamYaple | yea that a good place to start | 17:19 |
| SamYaple | also mysql was up there | 17:19 |
| sdake | if we add 2 gig of ram it won't make a big gdifference | 17:19 |
| SamYaple | it ate up like 1gb | 17:19 |
| sdake | (ram via swap) | 17:20 |
| SamYaple | it will if that is all active... | 17:20 |
| SamYaple | and in this case it is | 17:20 |
| sdake | well during deplooy it isn't | 17:20 |
| sdake | but during testing i guss it will be | 17:20 |
| SamYaple | it is | 17:20 |
| SamYaple | once the serivces start running its active | 17:21 |
| SamYaple | there isnt alot of stuff that could go to swap in the gate without severely impacting perfroamnce | 17:21 |
| SamYaple | i did all these tests | 17:21 |
| sdake | ok, wasn't ware you tested it | 17:21 |
| SamYaple | yea weve been over this | 17:21 |
| sdake | just a llittle dirty secret, alot of the vm runs in swap already on clodu proiders ) | 17:22 |
| SamYaple | thats not true | 17:22 |
| SamYaple | but ok | 17:22 |
| sdake | ram is always overcommitted | 17:22 |
| SamYaple | thats not the same thing as "vm runs in swap" | 17:22 |
| sdake | ok | 17:22 |
| sdake | so when the overcommit hits the limit, it goes to where? | 17:23 |
| SamYaple | and there is a reason drivers the like the baloon virtio driver exists | 17:23 |
| SamYaple | in the case of the virtio driver the host asks the guests to dump cache | 17:23 |
| sdake | i see | 17:23 |
| sdake | cool didn't know that | 17:23 |
| SamYaple | thats the whole idea. the host doesnt know whats important, but the guest does | 17:24 |
| sdake | so the host never puts stuff in swap on overcommit? | 17:24 |
| SamYaple | i didnt say that | 17:24 |
| sdake | SamYaple when you build ubuntu make sure to build with --no-cche please | 17:25 |
| sdake | britthouser you have to admit building and testing both centos and ubuntu in 25 minutes is pretty hot :) | 17:26 |
| sdake | testing/deploying | 17:26 |
| SamYaple | i never build with cache | 17:27 |
| sdake | oh | 17:27 |
| sdake | o kcool then | 17:27 |
| sdake | samyaple in tox.ini which is the right testenv - enabling oracle linux in the gate | 17:29 |
| sdake | build-ubuntu-source? | 17:29 |
| sdake | deploy-centos-binary? | 17:29 |
| SamYaple | what do you mean? | 17:29 |
| sdake | [testenv:build-images-centos-binary] | 17:29 |
| sdake | there is that | 17:29 |
| SamYaple | you know oracle is building with more than just raw kolla right? | 17:29 |
| SamYaple | i dont know that oraclelinux works just vaniliaa | 17:30 |
| sdake | [testenv:build-centos-binary] | 17:30 |
| sdake | upstream oracle works - i've tested it locally | 17:30 |
| sdake | which one of those tesetenvs is the right one to copy | 17:31 |
| SamYaple | any of them | 17:31 |
| sdake | you got rid of build-images right? | 17:31 |
| SamYaple | oh thats what you are asking | 17:31 |
| SamYaple | https://review.openstack.org/#/c/245096/ | 17:31 |
| sdake | thanks | 17:32 |
| sdake | boy our queue is backed up | 17:32 |
| SamYaple | ive reviewed everything.... | 17:32 |
| sdake | ya i've been slacking | 17:33 |
| sdake | i wasn't suggeting you were | 17:33 |
| sdake | so lights out on mariadb | 17:34 |
| sdake | is that a docs thing? | 17:34 |
| sdake | or is there any way to atuomate that | 17:34 |
| sdake | and lights out on rabbitmq, - is solved by clusterer? | 17:34 |
| SamYaple | it can be automated | 17:34 |
| SamYaple | to a degree | 17:35 |
| SamYaple | it will be a seperate special run playbook | 17:35 |
| SamYaple | dont have the cycles | 17:35 |
| sdake | any chance ou can document how to do it - i'll tackle it | 17:35 |
| sdake | or point me at some docs | 17:35 |
| sdake | or throw me some kind of bone | 17:35 |
| SamYaple | do you know how to recover from galera outage? both a controlled shutdown or a hardpoweroff? | 17:36 |
| sdake | if its a speical run playbook hw would it work since mariadb is marked to always restart? | 17:36 |
| sdake | no idea | 17:36 |
| SamYaple | i dont know how to help you then | 17:37 |
| SamYaple | there are docs describing it, but the implementation in kolla will be trickyy | 17:37 |
| sdake | ok well i'll hunt for docs on google later | 17:37 |
| sdake | my first question though is the restart=always? | 17:37 |
| SamYaple | what about it | 17:37 |
| sdake | are you suggesting we remove that for mariadb t odeal with lights out? | 17:37 |
| SamYaple | no | 17:38 |
| SamYaple | when did i say that/ | 17:38 |
| sdake | you didn't | 17:38 |
| SamYaple | oh ok | 17:38 |
| sdake | so if the mariadb process restarts, how does it not cause damage? | 17:38 |
| SamYaple | galera | 17:38 |
| sdake | galera | 17:38 |
| sdake | i guess i dont understand what the problem is with lights out that would require removing restart=always | 17:39 |
| sdake | would/wiouldn't | 17:39 |
| SamYaple | its my weekend i dont feel like explaining galera or quorums | 17:40 |
| sdake | after a restart, is galera in some sort of degraded operational state where it wont provide service but is running? | 17:40 |
| SamYaple | i will handle this if yo uwant to focus on something else | 17:40 |
| sdake | ok fair enough | 17:40 |
| SamYaple | there are other things needed for 1.1 that i am not planning on doing right now | 17:40 |
| sdake | SamYaple dealing with lights out is our highest development priority | 17:40 |
| sdake | upgrades being #2 | 17:40 |
| SamYaple | right. ill handle mariadb | 17:40 |
| SamYaple | either way we arent tagging 1.1 until all work is done | 17:41 |
| SamYaple | so there are other works to be done | 17:41 |
| sdake | could you enumerate | 17:41 |
| sdake | so we are on same page | 17:41 |
| SamYaple | check the etherpads | 17:41 |
| sdake | i'd be happy with lights out and upgrade | 17:41 |
| SamYaple | correct me if im wrong, but lights out is only an issue with galera | 17:41 |
| sdake | was this the upgrade ehterpad session? | 17:41 |
| SamYaple | i think so | 17:41 |
| sdake | SamYaple yes i htink so | 17:42 |
| sdake | ok let me check the pad | 17:42 |
| sdake | ok so we arent doing upgrades of mariadb exceept for major versions | 17:43 |
| sdake | so we dont have to solve that for 1.1 | 17:43 |
| SamYaple | did you pin centos? | 17:43 |
| SamYaple | i dont think ive pinned ubuntu | 17:43 |
| SamYaple | i dont remember pretty sure that was on the list | 17:44 |
| sdake | CentOS - Kolla has pinned to hammer, but not to a specific z version (sdake will look into this) | 17:44 |
| sdake | see I knew it was pinned :) sort of | 17:44 |
| sdake | mariadb is not pinned | 17:44 |
| SamYaple | thats ceph | 17:44 |
| sdake | atlaest on centos | 17:44 |
| SamYaple | not mariadb | 17:45 |
| sdake | Liberty dot-Y release: | 17:46 |
| sdake | ovsdb data container | 17:46 |
| sdake | rabbit cluster | 17:46 |
| sdake | keepalived no preempt | 17:46 |
| sdake | pin ceph to specific release for centos | 17:46 |
| sdake | Kolla needs to know what version is currently running (handled by /etc/kolla/release) | 17:46 |
| sdake | sorry for spam ;) | 17:46 |
| sdake | did you have a patch in the works for /etc/kolla/releae? | 17:46 |
| SamYaple | yea its a WIP just showing that i plan to drop te version at the very end | 17:47 |
| sdake | well what i'll do right now is put all thee dot-y release features sin as blueprints | 17:47 |
| sdake | since they are not bllueprintted | 17:48 |
| SamYaple | thats what ive been asking for a while. so +2 from me on that idea | 17:48 |
| sdake | i am not convinced we can pin ceph to as pecific version | 17:48 |
| sdake | sorry dude, like i said, i dont handle 15 hour time travel well | 17:48 |
| SamYaple | i do in ubuntu | 17:48 |
| SamYaple | and youll have to handle it | 17:48 |
| SamYaple | im not doing interrelseae ceph upgrades | 17:48 |
| sdake | i can pin it | 17:48 |
| sdake | but i they rdelete the rpm we are fucked ;) | 17:49 |
| SamYaple | like i said, youll have to figure it out | 17:49 |
| sdake | i'll reach out to some contacts there and seee if they can put up a version that wont be deleted | 17:49 |
| SamYaple | ceph is complicated to upgrade and im not doing it mid release | 17:49 |
| sdake | agreed 100% | 17:49 |
| sdake | the only ting I wantot upgrade midrelease is openstack services tyemselve | 17:49 |
| sdake | major release different story | 17:50 |
| sdake | so you will take on the lights out thing then? | 17:50 |
| sdake | if so i'll assign that blueprint to you | 17:50 |
| SamYaple | holy crap! cloud-archive guys put up a mitaka repo! | 17:51 |
| SamYaple | about time! | 17:51 |
| sdake | oh snap | 17:51 |
| sdake | are eyou serious so its like master? | 17:51 |
| sdake | damn we need to integrate with that! | 17:51 |
| sdake | rather its like delorean? | 17:51 |
| SamYaple | lol no its not master | 17:51 |
| SamYaple | not like deloread | 17:52 |
| * sdake confused | 17:52 | |
| sdake | used | 17:52 |
| sdake | and abused | 17:52 |
| SamYaple | they have the mitaka repo, but its just the liberty packages | 17:52 |
| SamYaple | they always start by cloning the old repo | 17:52 |
| SamYaple | normally they wait until 1 week before openstack release though | 17:52 |
| SamYaple | so maybe they are getting thier act together | 17:52 |
| SamYaple | maybe they will try to start publishing master is my point | 17:53 |
| sdake | right | 17:53 |
| sdake | its actually a pretty rough job | 17:53 |
| sdake | redhat has 5 people working on it | 17:53 |
| sdake | what is keepalive no preempt | 18:10 |
| sdake | SamYaple when you make a blueprint cn you se series goal please | 18:30 |
| sdake | otherwise it wont show up properly in the tracker | 18:31 |
| sdake | wow 56 blueprints | 18:32 |
| * sdake groans | 18:32 | |
| *** jasonsb has joined #kolla | 18:37 | |
| *** dims has joined #kolla | 18:39 | |
| *** dims has quit IRC | 18:44 | |
| sdake | SamYaple what do you think these guys are doing here | 18:55 |
| sdake | https://blueprints.launchpad.net/kolla/+spec/external-interface-bind | 18:55 |
| sdake | is 192.168.27.102 and 192.168.121.29 on the same box? | 18:56 |
| britthouser | sdake: no pre-empt is when a previously primary keepalived container comes back up from a failure, it won't take over primary again, causing unnecessary failover. | 19:05 |
| sdake | ya the bluepritn i wrote says that does it not | 19:05 |
| sdake | if not, click the yellow button and edit it | 19:05 |
| sdake | if you dont hae permissions ask me an i'll add you | 19:06 |
| britthouser | yeah taht sounds good | 19:06 |
| sdake | is the blueprint correct? | 19:09 |
| sdake | britthouser here is the blueprint i filed https://blueprints.launchpad.net/kolla/+spec/keepalived-no-preempt | 19:14 |
| *** bmace has quit IRC | 19:16 | |
| *** bmace has joined #kolla | 19:16 | |
| openstackgerrit | Merged openstack/kolla: Add usermod to Zaqar container https://review.openstack.org/245436 | 19:25 |
| *** sdake_ has joined #kolla | 19:29 | |
| *** sdake has quit IRC | 19:29 | |
| britthouser | Yeah bluerpint looks good | 19:35 |
| sdake_ | i get so many job requests | 19:36 |
| sdake_ | atleat 2-3 a day | 19:36 |
| sdake_ | but the ones that really get me are "come wor for us as a senior software engineer" | 19:36 |
| sdake_ | jesus dude, can you see I'm a PE at cisco? | 19:36 |
| sdake_ | its not like ciso hands out pe titles like candy | 19:37 |
| britthouser | lol | 19:37 |
| britthouser | I think I could take the keepalive blueprint. How do I sign up for that? | 19:37 |
| sdake_ | even some of the "tech leads" ive run into are formidable | 19:37 |
| britthouser | yeah each org has their own definition of what it takes to become tech lead. | 19:38 |
| britthouser | My last org it was like be active part of IETF | 19:38 |
| sdake_ | by tech lead i mean technical leader | 19:38 |
| britthouser | and then I've run into some tech leads that make me scratch my head. | 19:38 |
| sdake_ | i like cisco's teechnical ladder | 19:39 |
| sdake_ | in some respects i wish I had always worked at a company with a real technical ladder | 19:39 |
| sdake_ | because everywhere i've worked since montaviista, i've been at the logical top of the ladder | 19:39 |
| sdake_ | there are 84 des at cisco | 19:39 |
| sdake_ | so, pretty much, never going to happen | 19:39 |
| britthouser | Yeah you've still got DE and then fellow to go | 19:40 |
| sdake_ | britt houser yu assign yourself to it | 19:40 |
| sdake_ | i dont have hope of de | 19:40 |
| sdake_ | it is posssible | 19:40 |
| sdake_ | but i need alot more internal influence f or that to happen | 19:40 |
| sdake_ | thatwill take years | 19:40 |
| britthouser | ok maybe I'm just dense, but I dno't see a button for assign? | 19:40 |
| sdake_ | fortunately cisco highly values external influence :) | 19:40 |
| sdake_ | assignee you dont see a yellow button? | 19:41 |
| britthouser | no | 19:41 |
| britthouser | do I need to register myself as a member of kolla or something? | 19:42 |
| sdake_ | i just added you to olla drivers group | 19:42 |
| *** suro-patz has joined #kolla | 19:43 | |
| sdake_ | you should be able tod o your thing now | 19:43 |
| sdake_ | suro-patz yo bro how goes it | 19:43 |
| sdake_ | britthouser cuple notes about kolla-drivers group | 19:43 |
| sdake_ | this gives you nearly complete control over the launchpad tracker | 19:43 |
| britthouser | that let me assign myself. | 19:43 |
| sdake_ | yo ucan do everything but delete things | 19:43 |
| * britthouser feels the burden | 19:43 | |
| sdake_ | now one of the responsibilities of the kolla-drivers team is to triage bugs | 19:44 |
| sdake_ | you can triage bugs or chose not to do so | 19:44 |
| sdake_ | its really up to you | 19:44 |
| sdake_ | but if you triage a bug please do it correctly | 19:44 |
| sdake_ | can you spend 3 minutes with me so I can show yu how | 19:44 |
| *** sacharya has joined #kolla | 19:45 | |
| sdake_ | https://bugs.launchpad.net/kolla/+bug/1514871 | 19:45 |
| openstack | Launchpad bug 1514871 in kolla "Vagrantfile using libvirt doesn't work behind http proxy" [Undecided,New] | 19:45 |
| sdake_ | open that up | 19:45 |
| britthouser | yes.ok | 19:45 |
| sdake_ | ok there are 4 fields, they ALL must be assigned | 19:46 |
| sdake_ | or I go crazy finding lost bugs | 19:46 |
| sdake_ | sorry 3 ust tbe assigned | 19:46 |
| sdake_ | the asignee field can be left unassigned | 19:46 |
| sdake_ | so go ahead and triage the bug | 19:46 |
| sdake_ | and i'll coach you where ou go wrong | 19:46 |
| britthouser | importance: low? | 19:47 |
| sdake_ | do what you think is right | 19:47 |
| sdake_ | i'll offer suggetions ater you finish | 19:47 |
| britthouser | ok | 19:47 |
| sdake_ | i wont be aroun to ask questions when you triage | 19:47 |
| sdake_ | you hve to trust your judgement ;) | 19:47 |
| britthouser | Ok...I filled in three fields. | 19:48 |
| sdake_ | good | 19:49 |
| sdake_ | which phase if mitaka are we in | 19:49 |
| britthouser | 1 | 19:49 |
| sdake_ | which phase did you set | 19:49 |
| britthouser | 3 | 19:49 |
| sdake_ | so you wouldn't know this unless you could read my mind | 19:49 |
| sdake_ | which is entirely possible ;) | 19:49 |
| britthouser | you want all bugs fixed in the phase they are filed? | 19:49 |
| sdake_ | but we put all bugs in the current release | 19:49 |
| *** sacharya has quit IRC | 19:50 | |
| britthouser | ok easy enough | 19:50 |
| sdake_ | and if they dont finish we roll them | 19:50 |
| sdake_ | we have bugs from 9 months ago in mitaka-1 ;) | 19:50 |
| sdake_ | that way everyone can find them at one place | 19:50 |
| sdake_ | and it improves bug throughput | 19:50 |
| sdake_ | the other responsibility of the kolla-drivers team is to triage blueprints | 19:50 |
| sdake_ | but I prefer the core revers do that work | 19:50 |
| britthouser | sure that makes sense | 19:51 |
| sdake_ | to find bugs to triage | 19:51 |
| sdake_ | when your bored | 19:51 |
| sdake_ | click bugs | 19:51 |
| sdake_ | click status | 19:51 |
| sdake_ | which new bug looks out of place? | 19:51 |
| britthouser | t1469891 ? | 19:52 |
| sdake_ | right | 19:52 |
| sdake_ | its half-triaged | 19:52 |
| sdake_ | this is not a prolem because they left it in the new state thank god | 19:52 |
| sdake_ | if someone changes it from the new state and doesn't se tthe target release, the bug gets lost | 19:53 |
| sdake_ | i dontknow how to find them | 19:53 |
| sdake_ | so pleae, set all fields :) | 19:53 |
| britthouser | So the 3 must-have fields are: status, importance, milestone? | 19:53 |
| sdake_ | right | 19:53 |
| sdake_ | now if you confirm a bug | 19:53 |
| sdake_ | set confirm and re-set priority | 19:53 |
| sdake_ | after confirming a bug, the priority may hae changed | 19:53 |
| britthouser | confirm means I reproduced it myself? | 19:54 |
| sdake_ | like you guess when you triage | 19:54 |
| sdake_ | but when you c onfirm you atulally now | 19:54 |
| sdake_ | right | 19:54 |
| sdake_ | this is not reequired for triage | 19:54 |
| britthouser | makes sense | 19:54 |
| sdake_ | triage and confirmation are two separate things | 19:54 |
| sdake_ | dont feel like you have ot confirm a bug | 19:54 |
| britthouser | ok...so does that look better? | 19:54 |
| sdake_ | but if you do see a bug in the tracker, and you dupicate it, confirm it | 19:54 |
| sdake_ | so priority | 19:55 |
| sdake_ | a service is completely undeployable | 19:55 |
| sdake_ | people use or should be using the priority field to select which bugs to fix first | 19:56 |
| sdake_ | perosnally i would set to critical | 19:57 |
| britthouser | yeah do you have example criteria for each level? | 19:57 |
| sdake_ | unfortunately its subjective | 19:57 |
| sdake_ | wishlist = feature | 19:57 |
| britthouser | ok | 19:57 |
| sdake_ | that issn't worthy of a blueprint | 19:57 |
| sdake_ | that is set in stone | 19:57 |
| sdake_ | low = who cares | 19:58 |
| sdake_ | medium = ya its busted but doesn't totally damage the system | 19:58 |
| sdake_ | high = damages the system but still operational | 19:58 |
| sdake_ | critical = service not running | 19:58 |
| *** sdake_ is now known as sdake | 19:58 | |
| sdake | now just a heads up, the kolla-drivers team is different then the kolla-core team | 19:59 |
| sdake | some projects treat them one in the same but because f how launchpad permissions work i can't do that | 19:59 |
| sdake | kolla-core has more responsibilities | 19:59 |
| britthouser | Ok. | 20:00 |
| sdake | if you have desire to be a part of that team, I'd recommend to start reviewing patches | 20:00 |
| britthouser | I have desire, I'm hoping that after summit mgmt will support that. =) | 20:01 |
| sdake | well who knows | 20:01 |
| britthouser | yeah =) | 20:01 |
| britthouser | off to pick up b-day cake | 20:01 |
| britthouser | catchya after a bit | 20:02 |
| sdake | enjoy | 20:02 |
| *** sdake_ has joined #kolla | 20:03 | |
| *** sdake has quit IRC | 20:06 | |
| *** vilobhmm has joined #kolla | 20:14 | |
| sdake_ | samyaple can't tag liberty with this problem https://bugs.launchpad.net/kolla/+bug/1516306 | 20:59 |
| openstack | Launchpad bug 1516306 in kolla "neutron-server restart loop on stable/liberty" [Critical,Triaged] | 20:59 |
| sdake_ | neutron.conf:syslog_log_facility = LOG_LOCAL0 | 21:01 |
| sdake_ | https://bugs.launchpad.net/oslo.log/+bug/1499620 | 21:12 |
| openstack | Launchpad bug 1499620 in oslo.log "Unintended assignment of "syslog"" [High,Fix released] - Assigned to Masaki Matsushita (mmasaki) | 21:12 |
| *** vilobhmm has quit IRC | 21:28 | |
| SamYaple | sdake_: https://github.com/openstack/kolla/blob/stable/liberty/docker/openstack-base/Dockerfile.j2#L81 | 21:45 |
| SamYaple | there is a fix in place for that | 21:45 |
| SamYaple | oh right | 21:45 |
| SamYaple | you use centos | 21:46 |
| SamYaple | when is that packaging not going to be broken and shitty | 21:46 |
| sdake_ | ya | 21:46 |
| sdake_ | so rdo guys have said they are t-2 weeks from having a fully packstack based AIO gate | 21:47 |
| sdake_ | and our packages will come form tht gate | 21:47 |
| SamYaple | dont know what to tell you. dont hold up the tagging because binary entos is broken | 21:47 |
| sdake_ | they have also added a requirements.txt to rpm dependencies validation step in their gating | 21:48 |
| SamYaple | wont matter, requirements.txt has a broken version | 21:48 |
| sdake_ | we need to hunt a core reviewer down to approve two patches | 21:48 |
| sdake_ | i have one ready to go | 21:48 |
| sdake_ | and i assume you will approve them as well | 21:48 |
| SamYaple | for what | 21:49 |
| sdake_ | http://ur1.ca/o9rf9 -> http://paste.fedoraproject.org/290573/37764144 | 21:49 |
| sdake_ | cinder is also broken | 21:49 |
| SamYaple | all logging is broken | 21:50 |
| sdake_ | the reaon this happened is because we went from the master upstream repo to the liberty one | 21:50 |
| SamYaple | how is that going to fix it when all the other stuff is just going to pull in the package and overwrite the pipp one | 21:50 |
| sdake_ | i get logging is not workign correctly but this problem causes neutron not to work | 21:50 |
| sdake_ | still testing, I think I had meant to install oslo.log ahead of the pip install | 21:51 |
| sdake_ | jusst finished a deploy and checking if it worked moment | 21:52 |
| SamYaple | you should have probably tested that master to liberty repo change patch | 21:52 |
| sdake_ | y no beuno | 21:52 |
| sdake_ | your right i should have | 21:52 |
| sdake_ | bad move on my part | 21:52 |
| SamYaple | i checked the gate, but the gate was fine | 21:52 |
| sdake_ | i did as well | 21:52 |
| sdake_ | we trust our gate too much | 21:53 |
| sdake_ | i want to get to poit where we CAN trust our gate | 21:53 |
| SamYaple | i never trust it | 21:53 |
| SamYaple | i dont have centos | 21:53 |
| sdake_ | but with your hard work the gate has improved dramatically | 21:53 |
| sdake_ | yes like you said should have been tested | 21:53 |
| sdake_ | especially since stable change | 21:53 |
| sdake_ | but was not | 21:53 |
| sdake_ | i trusst the gate much more now then i di din the pasat | 21:54 |
| sdake_ | ithnk the core team ha finally stopped approving patches prior to gat epassage | 21:54 |
| sdake_ | \o/ | 21:54 |
| sdake_ | worst case we can tag monday | 21:57 |
| sdake_ | unless we can hunt down someone on hthe weekend | 21:57 |
| sdake_ | perhaps harmw is about | 21:57 |
| sdake_ | sam besidess thevarious upgrade blueprints, how do the essentials look here https://launchpad.net/kolla/+milestone/mitaka-1 | 21:58 |
| *** dims has joined #kolla | 22:07 | |
| *** sdake_ has quit IRC | 22:12 | |
| *** sacharya has joined #kolla | 22:19 | |
| *** sacharya has quit IRC | 22:23 | |
| *** suro-patz has quit IRC | 22:25 | |
| *** tzn has joined #kolla | 22:35 | |
| *** sdake has joined #kolla | 22:41 | |
| *** tzn has quit IRC | 22:41 | |
| *** cemmason1 is now known as cemmason | 23:00 | |
| *** sdake has quit IRC | 23:14 | |
| *** dims has quit IRC | 23:47 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!