| cgoncalves | a follow-up to the meeting, I intent to spend time tomorrow understanding why some tests fail ocasionally on the centos job | 00:17 |
|---|---|---|
| cgoncalves | xgerman's amp log patches would help a lot flagging what's going on | 00:17 |
| xgerman | You can always +2 | 00:25 |
| cgoncalves | xgerman, not yet, the amphora log file is not being copied | 00:29 |
| xgerman | It is for legacy jobs | 00:30 |
| xgerman | Added patch for the other ones, too | 00:30 |
| cgoncalves | xgerman, no, I was looking already in to https://review.openstack.org/#/c/626406/ | 00:33 |
| cgoncalves | http://logs.openstack.org/06/626406/1/check/octavia-v2-dsvm-noop-py2-api-stable-rocky/3c5deaf/controller/logs/ | 00:33 |
| cgoncalves | no octavia-amphora.log file | 00:33 |
| xgerman | :-( | 00:34 |
| cgoncalves | wrong link, same thing -- http://logs.openstack.org/06/626406/1/check/octavia-v2-dsvm-py2-scenario-stable-rocky/a9e34bf/controller/logs/ | 00:34 |
| xgerman | Ok. Saw one for the legacy job and did same for the others | 00:34 |
| xgerman | Anyhow will check on it later | 00:35 |
| *** celebdor has quit IRC | 00:36 | |
| cgoncalves | is there a way to check which OS runs on the host machine in CI? (the server that hosts devstack VMs) | 00:45 |
| cgoncalves | wondering if RAX hypervisors run Ubuntu 16.04. https://bugs.launchpad.net/ubuntu/+source/linux/+bug/1783952 could be the result of centos job failing | 00:46 |
| openstack | Launchpad bug 1783952 in linux (Ubuntu) "Xubuntu 16.04 L0 hypervisor kernel panics L1 CentOS 7.5.1804 guest" [High,Confirmed] | 00:46 |
| cgoncalves | I see loads of qemu-kvm errors | 00:47 |
| cgoncalves | seems to only happen on rax and limestone nodes | 01:02 |
| logan- | cgoncalves: limestone nodes are all running 16.04 w/ hwe kernel 4.15.0-42.45~16.04.1 | 01:04 |
| cgoncalves | inap and ovh are okay | 01:04 |
| rm_work | is that also why we see periodic grenade failures? | 01:07 |
| cgoncalves | logan-, I'm not 100% sure that kernel is buggy but reading the launchpad I linked, I'd tend to say yes | 01:08 |
| cgoncalves | rm_work, periodic grenade failures? I haven't noticed that recently | 01:09 |
| rm_work | https://review.openstack.org/#/c/626132/ ? | 01:10 |
| rm_work | ah was it a timeout? | 01:10 |
| cgoncalves | rm_work, ah, rechecked that one. timeout | 01:10 |
| cgoncalves | http://zuul.openstack.org/builds?job_name=octavia-grenade | 01:10 |
| cgoncalves | seems stable enough | 01:10 |
| rm_work | when i looked earlier it didn't have that comment yet, and usually it SAYS "TIMEOUT" | 01:11 |
| cgoncalves | true | 01:11 |
| cgoncalves | xgerman, ha ah! http://logs.openstack.org/06/626406/1/check/octavia-v2-dsvm-scenario/da19664/controller/logs/octavia-amphora_log.txt | 01:17 |
| cgoncalves | hmm I mean, sure, stupid me. I was checking before jobs of stable releases which naturally do not have amp logging... | 01:18 |
| sapd1_ | https://review.openstack.org/#/c/624835/ Maybe We will lost log. If we have millions lines of log. | 01:26 |
| sapd1_ | Because rsyslog server do not perform caching. | 01:27 |
| *** sapd1 has joined #openstack-lbaas | 01:30 | |
| xgerman | sapd1 yes, I am a bit disappointed with Rsyslog HA attributes | 01:36 |
| sapd1 | xgerman: I have implemented logging for amphora and use filebeat, then filebeat will ship log to redis cluster. | 01:37 |
| xgerman | Yeah, we use filebeat internally, too, but might be too heavy for devstack/gate | 01:39 |
| *** abaindur has joined #openstack-lbaas | 01:41 | |
| *** abaindur has quit IRC | 01:42 | |
| *** abaindur has joined #openstack-lbaas | 01:43 | |
| rm_work | what about syslog-ng? | 01:47 |
| rm_work | rsyslog is absolute garbage | 01:47 |
| sapd1 | rm_work: I think we need a caching software such as redis to prevent loss of log. | 01:50 |
| *** sapd1 has quit IRC | 01:54 | |
| *** sapd1_x has joined #openstack-lbaas | 01:54 | |
| rm_work | yeah so, allowing output of logs from the amp in a configurable format (syslog, kafka, splunk, etc?) and then the user providing the endpoint details, would mean they can do whatever they want for HA and such | 01:55 |
| rm_work | sapd1_x: ^^ | 01:55 |
| rm_work | and that would be an API thing | 01:55 |
| rm_work | i still need to look at what exactly xgerman is doing, whether it's that or something else | 01:56 |
| sapd1_x | rm_work: With kafka and splunk we need install agent inside amp. | 01:57 |
| rm_work | yeah, well like... fluentd does output in all of those formats, for example | 01:57 |
| sapd1_x | great. We can use filebeat or fluentd. so we can config output driver for agent. | 01:58 |
| rm_work | actually, syslog-ng has kafka and redis output, for example | 01:58 |
| sapd1_x | great | 01:58 |
| rm_work | so honestly, maybe we just use syslog-ng and leave it at that | 01:58 |
| sapd1_x | as xgerman said, it might be heavy for gate. | 01:59 |
| rm_work | syslog-ng? | 01:59 |
| rm_work | i don't think so | 01:59 |
| sapd1_x | syslog-ng can ship log to redis? | 01:59 |
| rm_work | yes | 01:59 |
| sapd1_x | great | 01:59 |
| rm_work | https://www.syslog-ng.com/technical-documents/doc/syslog-ng-open-source-edition/3.16/administration-guide/27#TOPIC-956486 | 02:00 |
| rm_work | huge list of output drivers available | 02:00 |
| rm_work | can do direct to graphite or elastic | 02:00 |
| sapd1_x | I would like to submit a presentation for next Summit. :D | 02:01 |
| rm_work | i'd co-talk on that with you if you want :P | 02:01 |
| rm_work | tho xgerman is in the middle of actually DOING this, it seems, lol | 02:02 |
| sapd1_x | about logging and monitoring. But I have customized octavia for internal. Is it ok. | 02:02 |
| rm_work | yeah, really need to get that upstreamed | 02:03 |
| rm_work | talks about internal stuff are ... <_< | 02:03 |
| sapd1_x | I have sent the presentation to johnsom :D | 02:04 |
| johnsom | I agree that syslog-ng is a better choice | 02:17 |
| xgerman | I was wondering that myself… but please comment on the patch | 02:24 |
| xgerman | I can totally switch to syslog-ng :-) | 02:24 |
| xgerman | but wanted to start somewhere… | 02:24 |
| xgerman | also everything talks syslog… and rsyslog was already there | 02:25 |
| cgoncalves | some popular distros do not ship syslog-ng... | 02:27 |
| xgerman | yeah, I am not into supporting a million formats - as I said everyhting talks syslog. And if rsyslog can survive/buffer an outage of the system it sends to I am good | 02:28 |
| johnsom | Only the good ones.... | 02:29 |
| johnsom | cgoncalves Let's move the conversation here. So why is centos running so much slower? See this patch: https://review.openstack.org/#/c/626227/ | 02:30 |
| johnsom | 1:30 for ubuntu, 2:31 for centos | 02:30 |
| johnsom | That is likely why you are getting gate timeouts. Plus RAX hosts are slower than the others, so that probably doesn't help | 02:30 |
| *** rcernin has quit IRC | 02:31 | |
| cgoncalves | good question | 02:31 |
| johnsom | One recent change I see is all of the nodepool instances are bionic now | 02:32 |
| openstackgerrit | Carlos Goncalves proposed openstack/octavia-tempest-plugin master: DNM: test centos-7 job with default nodeset (ubuntu) https://review.openstack.org/626429 | 02:35 |
| cgoncalves | the centos7 job doesn't use default nodeset (ubuntu) but centos | 02:37 |
| cgoncalves | ^ might give us an indication of whether the problem lays in controller or amphora | 02:38 |
| *** psachin has joined #openstack-lbaas | 02:42 | |
| * bzhao__ johnsom: sorry to disturb your conversation. ;-) . I post the new patch yesterday night for http health check feature. But I didn't test the UT in py35. So I found the gap between py27 and py35, as the encode 'string_escape' doesn't exist in py35. Then I change from 'string_escape' to 'unicode_escape', the result is that py27 works, but not py35. Also notes that, I test with the way which we discussed a few | 02:42 | |
| * bzhao__ days ago, none of them can print/write to file with the correct '\n' sting. Then only the way in https://review.openstack.org/#/c/626183/1/octavia/common/jinja/haproxy/templates/macros.j2 .. | 02:42 | |
| bzhao__ | The current test in py35 seems not work correctly with unicode_escape. | 02:43 |
| bzhao__ | https://usercontent.irccloud-cdn.com/file/I0dHTUl5/image.png | 02:43 |
| bzhao__ | py35 result. But it works OK in py27. https://usercontent.irccloud-cdn.com/file/IjvvvDEP/image.png | 02:44 |
| johnsom | bzhao__ since you broke that string out, you can replace .encode() with |safe | 02:44 |
| johnsom | We just can't use |safe on any string with user input | 02:45 |
| *** sapd1_x has quit IRC | 02:48 | |
| bzhao__ | Agree, But I change to safe, same it not change with correct '\\n', it process as a new line. https://usercontent.irccloud-cdn.com/file/nFeUY1H5/image.png | 02:50 |
| bzhao__ | https://usercontent.irccloud-cdn.com/file/UZG9Ayrh/image.png | 02:50 |
| johnsom | Really... hmmm try \\n with safe | 02:50 |
| *** rcernin has joined #openstack-lbaas | 02:55 | |
| bzhao__ | This is in py27.. Oh, I found py27 is string, py35 is byte type..Is that the issue? Ha, just guess. But I change '\r\n' to '\\r\\n' . I need to test with file writing. Please hold on. This seems work. ;-) | 02:55 |
| johnsom | Oh good. I need to run to dinner, so I will chat later | 02:57 |
| johnsom | or tomorrow | 02:57 |
| *** PagliaccisCloud has joined #openstack-lbaas | 02:59 | |
| bzhao__ | OK. Thank you Michael. And the Christmas is comming, Have a nice holiday, Merry Christmas!! Also Be happy, All Octavia guys. ;-) | 02:59 |
| *** hongbin has joined #openstack-lbaas | 03:22 | |
| openstackgerrit | ZhaoBo proposed openstack/octavia master: Support Host header inject for healthmonitor HTTP 1.1 health check https://review.openstack.org/626183 | 04:02 |
| openstackgerrit | ZhaoBo proposed openstack/python-octaviaclient master: Add new options to HealthMonitor CLI https://review.openstack.org/626190 | 04:04 |
| *** rcernin has quit IRC | 04:05 | |
| *** sapd1_x has joined #openstack-lbaas | 04:07 | |
| *** hongbin has quit IRC | 04:08 | |
| *** rcernin has joined #openstack-lbaas | 04:12 | |
| *** hongbin has joined #openstack-lbaas | 04:43 | |
| *** sapd1_x has quit IRC | 04:55 | |
| *** sapd1_x has joined #openstack-lbaas | 05:07 | |
| openstackgerrit | Michael Johnson proposed openstack/octavia master: Add amphora statistics to the admin API https://review.openstack.org/585031 | 05:18 |
| openstackgerrit | Michael Johnson proposed openstack/octavia master: Add amphora statistics to the admin API https://review.openstack.org/585031 | 05:22 |
| *** hongbin has quit IRC | 05:24 | |
| *** abaindur has quit IRC | 05:43 | |
| *** abaindur has joined #openstack-lbaas | 05:44 | |
| *** gcheresh has joined #openstack-lbaas | 06:04 | |
| openstackgerrit | ZhaoBo proposed openstack/octavia master: Support L7policy redirect http code https://review.openstack.org/625007 | 06:16 |
| *** annp has joined #openstack-lbaas | 06:48 | |
| *** velizarx has joined #openstack-lbaas | 06:51 | |
| *** rcernin has quit IRC | 06:56 | |
| *** ccamposr has joined #openstack-lbaas | 07:03 | |
| *** velizarx has quit IRC | 07:10 | |
| *** PagliaccisCloud has quit IRC | 07:18 | |
| *** pcaruana has joined #openstack-lbaas | 07:25 | |
| *** velizarx has joined #openstack-lbaas | 07:29 | |
| *** celebdor has joined #openstack-lbaas | 07:41 | |
| *** yboaron_ has joined #openstack-lbaas | 07:41 | |
| openstackgerrit | ZhaoBo proposed openstack/octavia master: Support Host header inject for healthmonitor HTTP 1.1 health check https://review.openstack.org/626183 | 07:41 |
| *** gcheresh has quit IRC | 07:57 | |
| *** gcheresh has joined #openstack-lbaas | 07:57 | |
| *** rpittau has joined #openstack-lbaas | 08:09 | |
| *** celebdor has quit IRC | 09:02 | |
| *** abaindur has quit IRC | 09:07 | |
| *** celebdor has joined #openstack-lbaas | 09:10 | |
| *** PagliaccisCloud has joined #openstack-lbaas | 09:16 | |
| *** brtknr has joined #openstack-lbaas | 09:37 | |
| *** PagliaccisCloud has quit IRC | 10:19 | |
| *** celebdor has quit IRC | 10:19 | |
| openstackgerrit | Carlos Goncalves proposed openstack/octavia-tempest-plugin master: DNM: test centos-7 job with default nodeset (ubuntu) https://review.openstack.org/626429 | 10:38 |
| *** Krast has quit IRC | 10:41 | |
| *** celebdor has joined #openstack-lbaas | 10:44 | |
| *** salmankhan has joined #openstack-lbaas | 10:46 | |
| *** salmankhan has quit IRC | 11:41 | |
| *** salmankhan has joined #openstack-lbaas | 11:42 | |
| *** salmankhan has quit IRC | 11:46 | |
| *** salmankhan has joined #openstack-lbaas | 11:50 | |
| openstackgerrit | Carlos Goncalves proposed openstack/octavia-tempest-plugin master: DNM: test centos-7 job with default nodeset (ubuntu) https://review.openstack.org/626429 | 12:22 |
| *** yboaron_ has quit IRC | 12:30 | |
| *** yboaron_ has joined #openstack-lbaas | 12:31 | |
| *** rpittau is now known as rpittau|lunch | 12:32 | |
| *** sapd1_x has quit IRC | 12:44 | |
| openstackgerrit | Carlos Goncalves proposed openstack/octavia master: Add amphora statistics to the admin API https://review.openstack.org/585031 | 12:45 |
| *** yboaron_ has quit IRC | 13:02 | |
| *** velizarx has quit IRC | 13:11 | |
| *** velizarx has joined #openstack-lbaas | 13:22 | |
| *** yboaron_ has joined #openstack-lbaas | 13:24 | |
| *** psachin has quit IRC | 13:27 | |
| *** salmankhan has quit IRC | 13:27 | |
| *** rpittau|lunch is now known as rpittau | 13:33 | |
| *** yboaron_ has quit IRC | 13:33 | |
| *** yboaron_ has joined #openstack-lbaas | 13:34 | |
| *** salmankhan has joined #openstack-lbaas | 14:22 | |
| *** Swami has joined #openstack-lbaas | 14:46 | |
| *** jackivanov has joined #openstack-lbaas | 14:58 | |
| *** yboaron_ has quit IRC | 15:00 | |
| *** yboaron_ has joined #openstack-lbaas | 15:11 | |
| *** rpittau has quit IRC | 15:26 | |
| *** PagliaccisCloud has joined #openstack-lbaas | 15:27 | |
| *** velizarx has quit IRC | 15:45 | |
| *** gcheresh has quit IRC | 15:45 | |
| *** yboaron_ has quit IRC | 16:02 | |
| *** PagliaccisCloud has quit IRC | 16:15 | |
| openstackgerrit | Michael Johnson proposed openstack/octavia master: Updates Octavia to support octavia-lib https://review.openstack.org/613709 | 16:20 |
| cgoncalves | any particular reason for defaulting # of threads/processes to None which defaults to # of CPU cores? | 16:40 |
| cgoncalves | futuristics spawns that many threads/processes (max_workers) and never releases when executors complete | 16:42 |
| johnsom | Well, in general having it based on the number of cores is the right answer | 16:43 |
| cgoncalves | on nodes with a few 10s of CPU cores, that's a lot of overhead memory | 16:43 |
| cgoncalves | threads/processes sit idle | 16:43 |
| cgoncalves | killing and spawning them every so often is also not good | 16:43 |
| cgoncalves | unfortunately futuristic lib doesn't have min_workers | 16:43 |
| johnsom | That is why there is a config setting..... Plus, if you install Octavia on that much hardware, you must have some heavy load.... grin | 16:43 |
| johnsom | Which process are you looking at? | 16:44 |
| cgoncalves | health manager | 16:44 |
| cgoncalves | on a 16-core system, I get 35 processes total | 16:44 |
| johnsom | Hmmm, I thought that was capped by default..... | 16:45 |
| cgoncalves | capped to what? | 16:45 |
| johnsom | Yeah, one pool for heartbeats, one for stats | 16:45 |
| cgoncalves | right, that's already 32 processes (2 * 16) right at init | 16:46 |
| johnsom | Yeah, plus some masters, etc. | 16:46 |
| johnsom | Ah, ok, yeah it is uncapped. | 16:46 |
| johnsom | failover is capped. | 16:46 |
| cgoncalves | 16 * 2 + 1x health listener + hm_health_check + master = 35 | 16:47 |
| johnsom | Yeah, so you will be able to scale to a bunch of amps.... | 16:47 |
| johnsom | nice | 16:47 |
| amuller | neutron also doesn't cap number of workers | 16:47 |
| amuller | we cap it in tripleo | 16:47 |
| johnsom | Yeah, I am pretty sure none of the services do. But we do provide configuration settings to allow you to do so. | 16:48 |
| cgoncalves | ok, I wanted to understand if there was any reason behind not capping pool executors to a lower number of workers | 16:49 |
| cgoncalves | amuller, yeah, we'll need to do it in puppet-octavia. Joe volunteered himself to work on it | 16:50 |
| cgoncalves | and then some basic tripleo-heat-templates work, easy | 16:51 |
| johnsom | More power Scotty! | 16:51 |
| johnsom | I guess I could have answered your first question with: "To Boldly Go Where No Man Has Gone Before..." | 16:53 |
| cgoncalves | "... and play safe to lower chances of getting customer tickets" | 16:56 |
| *** Krast has joined #openstack-lbaas | 17:14 | |
| *** ccamposr has quit IRC | 17:19 | |
| *** Swami has quit IRC | 17:30 | |
| *** dmellado has quit IRC | 17:31 | |
| *** celebdor has quit IRC | 17:58 | |
| openstackgerrit | Merged openstack/octavia master: Set same debug log level as per DevStack local.conf https://review.openstack.org/626132 | 18:15 |
| *** Swami has joined #openstack-lbaas | 18:18 | |
| *** ianychoi has quit IRC | 18:22 | |
| *** salmankhan has quit IRC | 18:27 | |
| *** ipv4depletion has joined #openstack-lbaas | 18:29 | |
| *** dmellado has joined #openstack-lbaas | 18:40 | |
| openstackgerrit | Michael Johnson proposed openstack/octavia master: Add flavor, flavor_profile table and their APIs https://review.openstack.org/486499 | 18:44 |
| openstackgerrit | Michael Johnson proposed openstack/octavia master: Add flavors/flavor_profile api-ref https://review.openstack.org/618322 | 18:49 |
| *** JudeCross has joined #openstack-lbaas | 18:56 | |
| *** ipv4depletion has quit IRC | 19:01 | |
| openstackgerrit | Michael Johnson proposed openstack/octavia master: Add flavor, flavor_profile table and their APIs https://review.openstack.org/486499 | 19:12 |
| *** amuller has quit IRC | 19:14 | |
| openstackgerrit | Michael Johnson proposed openstack/octavia master: Add flavors/flavor_profile api-ref https://review.openstack.org/618322 | 19:15 |
| *** abaindur has joined #openstack-lbaas | 19:45 | |
| *** jackivanov has quit IRC | 19:54 | |
| *** jackivanov has joined #openstack-lbaas | 19:54 | |
| *** pcaruana has quit IRC | 19:57 | |
| *** jackivanov has quit IRC | 19:59 | |
| *** jackivanov has joined #openstack-lbaas | 20:06 | |
| *** jackivanov has quit IRC | 20:09 | |
| *** abaindur has quit IRC | 20:09 | |
| *** jackivanov has joined #openstack-lbaas | 20:09 | |
| *** jackivanov has quit IRC | 20:14 | |
| *** jackivanov has joined #openstack-lbaas | 20:22 | |
| openstackgerrit | Andreas Jaeger proposed openstack/neutron-lbaas-dashboard master: Use template for lower-constraints https://review.openstack.org/626711 | 20:22 |
| *** jackivanov has quit IRC | 20:23 | |
| *** jackivanov has joined #openstack-lbaas | 20:24 | |
| *** jackivanov has quit IRC | 20:26 | |
| *** jackivanov has joined #openstack-lbaas | 20:27 | |
| *** jackivanov has quit IRC | 20:31 | |
| openstackgerrit | Michael Johnson proposed openstack/octavia master: Add flavor, flavor_profile table and their APIs https://review.openstack.org/486499 | 21:06 |
| johnsom | blah, alembic head moved form the tags patch | 21:06 |
| openstackgerrit | Michael Johnson proposed openstack/octavia master: Add flavors/flavor_profile api-ref https://review.openstack.org/618322 | 21:06 |
| openstackgerrit | Carlos Goncalves proposed openstack/python-octaviaclient master: Add failover an amphora client support https://review.openstack.org/532424 | 21:15 |
| cgoncalves | this ^ has been sitting for almost a year now. back then we discussed where to place amphora-specific commands in the openstack cli but nothing has changed since. I'm in favor of accepting "openstack loadbalancer amphora failover" at least for now | 21:17 |
| cgoncalves | I should probably open a story and attach it to the patch, unless there's one already | 21:18 |
| openstackgerrit | Carlos Goncalves proposed openstack/python-octaviaclient master: Add failover an amphora client support https://review.openstack.org/532424 | 21:23 |
| johnsom | Hmmm, could do some sort of provider namespace | 21:26 |
| *** PagliaccisCloud has joined #openstack-lbaas | 21:44 | |
| rm_work | hmmmm | 21:54 |
| rm_work | well | 21:54 |
| rm_work | we did already merge some commands i thought? | 21:54 |
| rm_work | or did it seriously sit there this whole time | 21:54 |
| johnsom | Yeah, I thought so too | 21:54 |
| rm_work | amphora stats? | 21:54 |
| rm_work | or no | 21:54 |
| rm_work | i guess no | 21:54 |
| rm_work | maybe we really didn't | 21:54 |
| rm_work | ah yeah | 21:55 |
| rm_work | loadbalancer amphora [list|show] | 21:55 |
| johnsom | openstack loadbalancer amphora list | 21:55 |
| johnsom | Contrary to some opinions, we do have docs... https://docs.openstack.org/python-octaviaclient/latest/cli/index.html#amphora | 21:55 |
| johnsom | grin | 21:55 |
| rm_work | so maybe it's ok to do "openstack loadbalancer <provider> <commands> | 21:55 |
| rm_work | " | 21:55 |
| rm_work | loadbalancer is our namespace, so we can pretty much control what goes in there | 21:56 |
| rm_work | and i doubt providers are going to collide with base commands | 21:56 |
| rm_work | though MAYBE the verbose "loadbalancer provider amphora list" might have been better? | 21:56 |
| rm_work | making a "provider" namespace | 21:56 |
| rm_work | that's a lot to type <_< | 21:56 |
| rm_work | are we going to have 'loadbalancer provider list' and 'loadbalancer flavor list'? | 21:57 |
| johnsom | yeah, also could do openstack provider, but then.... types, yuck | 21:57 |
| rm_work | are providers going to also merge their commands into the main client? | 21:57 |
| rm_work | or do we need a plugin system for our plugin | 21:57 |
| cgoncalves | we discussed this and even voted for | 21:57 |
| johnsom | Someone has already create a patch for that, but I haven't looked at it yet. | 21:58 |
| johnsom | providers can create their own OSC plugins | 21:58 |
| johnsom | If they want | 21:58 |
| cgoncalves | all I'm saying is that patch has been idle for a year for no reason because, as you both just said, we have "openstack loadbalance amphora [list|show]" | 21:59 |
| rm_work | hmmmm | 22:00 |
| cgoncalves | this patch wouldn't open a new precedence IMO | 22:00 |
| rm_work | yeah looking at my last comment.... | 22:00 |
| rm_work | i guess we failed at being consistent with our voting lol | 22:01 |
| cgoncalves | the other day I found myself typing "... amphora failover $id" | 22:01 |
| cgoncalves | meaning I, as a regular user, would have expected that command to exist and be just like that | 22:02 |
| johnsom | As a regular user that command will tell you "NO" | 22:02 |
| johnsom | grin | 22:02 |
| cgoncalves | right, and I got pissed and failed over the LB | 22:03 |
| johnsom | Can't run that as a regular user either | 22:03 |
| cgoncalves | I could have made a curl put request but I was lazy... and a regular user :) | 22:03 |
| cgoncalves | don't ruin the point I'm trying to make -_- | 22:04 |
| cgoncalves | s/regular user/admin user/ | 22:04 |
| johnsom | Grin. | 22:04 |
| johnsom | Well, we have 6 cores... | 22:04 |
| johnsom | Personally I think that ship has sailed since we merged the other commands. That doesn't mean we can't alias those into another namespace later. | 22:05 |
| cgoncalves | yep, so at least we could offering amp failover | 22:06 |
| cgoncalves | *could start | 22:06 |
| rm_work | k | 22:07 |
| rm_work | thats fine | 22:07 |
| *** ivve has quit IRC | 22:08 | |
| johnsom | This one makes sense, but ugh is long: openstack loadbalancer provider capability show <provider> | 22:09 |
| cgoncalves | keystrokes are still free | 22:16 |
| johnsom | Yeah, and tab completion works. It's not like that is a command you are going to run every day | 22:16 |
| *** ivve has joined #openstack-lbaas | 22:20 | |
| rm_work | cgoncalves: that is SO not true, lol | 22:24 |
| rm_work | also: you should see the alias section of my bashrc >_> | 22:25 |
| rm_work | it contains gems like: alias oslb="openstack loadbalancer" | 22:25 |
| rm_work | alias gl="git log --stat" | 22:26 |
| cgoncalves | what you save on keystrokes with aliases you spend on brain cycles to recollect all those | 22:26 |
| rm_work | lol no, muscle memory now | 22:26 |
| johnsom | Yeah, sadly my keystrokes are not free.... | 22:26 |
| rm_work | all my git aliases cause me trouble sometimes on other systems | 22:27 |
| cgoncalves | muscle has to be fed. feeding is expensive | 22:27 |
| rm_work | gs/gl/gc/gd | 22:27 |
| rm_work | gs happens to conflict with ghostscript <_< | 22:27 |
| johnsom | Yeah, I have clonetempest and such | 22:27 |
| rm_work | woah that is long | 22:27 |
| rm_work | shouldn't it be 'ct'? :P | 22:27 |
| cgoncalves | lol! | 22:27 |
| johnsom | well octavia is cloneme so shorter | 22:28 |
| rm_work | 'co' | 22:28 |
| cgoncalves | you're killing me xD | 22:28 |
| rm_work | 'ykm' | 22:29 |
| rm_work | sorry, 'ykmxd' | 22:29 |
| cgoncalves | 'kc' (kill carlos) is shorter | 22:29 |
| rm_work | true | 22:29 |
| rm_work | alias kc='kill -9 carlos' | 22:29 |
| cgoncalves | heh | 22:30 |
| *** rcernin has joined #openstack-lbaas | 22:50 | |
| openstackgerrit | Merged openstack/python-octaviaclient master: Add failover an amphora client support https://review.openstack.org/532424 | 23:16 |
| xgerman | rm_work: https://www.rsyslog.com/doc/master/configuration/modules/omkafka.html — makes me less interested in switching rsyslog out… but yeah, comment on patch. Also cgoncalves it looks like syslog-ng is in epel which might/might not work with CentOS/Rhel/etc. | 23:18 |
| rm_work | syslog-ng is available on far more platforms than rsyslog AFAIK <_< | 23:19 |
| rm_work | also, my experience using rsyslog is that it is literal trash | 23:19 |
| rm_work | it's rare that a piece of software makes me *angry*, but rsyslog managed | 23:21 |
| xgerman | I did soemthign with syslog-ng 4 year ago because rsyslog was trash… from the docs it seems it can do the things it lacked those days. | 23:21 |
| rm_work | my experience was that it would arbitrarily ignore valid configuration options, and do so silently, but only sometimes | 23:22 |
| colin- | oh how confusing, i guess the game warframe has an item in called octavia.. and it can be "amp"ed somehow heh | 23:23 |
| colin- | lead to some odd search results | 23:23 |
| rm_work | lol | 23:23 |
| xgerman | lol | 23:23 |
| rm_work | lol yes "Octavia Amp damage boost" | 23:23 |
| johnsom | We have influence beyond our knowledge.... | 23:23 |
| colin- | >Not many results contain ca. | 23:23 |
| colin- | you don't say... | 23:23 |
| xgerman | First a car, now. a game,... | 23:24 |
| rm_work | https://warframe.fandom.com/wiki/Amp_(Ability) | 23:24 |
| johnsom | There is a restaurant in San Francisco too. | 23:24 |
| rm_work | yeah, i didn't make it out there, i'm sad | 23:24 |
| johnsom | https://www.octavia-sf.com/ | 23:25 |
| johnsom | Is this good enough for when a flavor is disabled? "The selected flavor is not allowed in this deployment: 83cc7079-7936-474e-9a92-eb63294f2070" | 23:29 |
| johnsom | Or should I create a new exception? | 23:29 |
| openstackgerrit | Michael Johnson proposed openstack/octavia master: Adds flavor support to the amphora driver https://review.openstack.org/621323 | 23:39 |
| openstackgerrit | Michael Johnson proposed openstack/octavia master: Add provider driver capabilities API https://review.openstack.org/624190 | 23:41 |
| johnsom | Ok, I think the flavors code chain is in good shape. Now, docs and release note | 23:41 |
| rm_work | 🤯🤯🤯 | 23:46 |
| johnsom | That looks like a sheeple emoji here.... | 23:48 |
| rm_work | lol | 23:48 |
| rm_work | it's "head exploding" | 23:48 |
| johnsom | Yeah, hover reveals the true nature.... | 23:48 |
| rm_work | ah not sure if my message before that went through | 23:54 |
| johnsom | Nope, just the emoji | 23:55 |
| rm_work | it was: awesome, i love how '--limit' for 'openstack image list' is just ... ignored | 23:55 |
| johnsom | I just assumed you were excited about the flavors work finally getting implemented | 23:56 |
| *** JudeCross has quit IRC | 23:56 | |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!
