| *** amitry has joined #openstack-lbaas | 00:00 | |
| *** bar_ has quit IRC | 00:00 | |
| *** AlexeyAbashkin has quit IRC | 00:02 | |
| *** bar_ has joined #openstack-lbaas | 00:11 | |
| johnsom | Hmm, IRCcloud is acting strange | 00:29 |
|---|---|---|
| xgerman_ | okey | 00:29 |
| johnsom | Anyway, I just tested the QoS patch and it seems to work | 00:29 |
| xgerman_ | yeah, I did we ever get around instaling qos in devstack/ | 00:30 |
| *** tongl has quit IRC | 00:31 | |
| johnsom | I just followed the networking guide to enable it in neutron, then since it messes with the ovs interface types I had to create a new subnet to use for testing. | 00:31 |
| johnsom | The "don't I wish" virtual networking speed: 16.2 Gbits/sec , the "through an LB" speed: 7.79 Gbits/sec, the QoS enabled speed: 3.07 Mbits/sec | 00:32 |
| xgerman_ | yeah, I think if we support QoS we should enable it by default ibn our devstack | 00:32 |
| xgerman_ | maybe I need to create a story | 00:32 |
| johnsom | The one downside I see is if you enable it, you can't un-enable it. | 00:32 |
| xgerman_ | mmh, true | 00:33 |
| *** sshank has quit IRC | 00:33 | |
| johnsom | Our validation requires a uuid | 00:33 |
| xgerman_ | it’s not like we instal BBQ every time so you can do SSL | 00:34 |
| xgerman_ | so that’s prob, fine | 00:34 |
| johnsom | Ah, you can pass in null via the API and remove it | 00:35 |
| *** fnaval_ has joined #openstack-lbaas | 00:35 | |
| *** fnaval has quit IRC | 00:35 | |
| xgerman_ | I also would like to see that QoS featured in our cookbook | 00:35 |
| *** fnaval_ has quit IRC | 00:35 | |
| johnsom | Comment away | 00:36 |
| xgerman_ | not sure if that is more me making tasks | 00:36 |
| rm_work | actually our gates DO all install Barbican | 00:37 |
| johnsom | Yeah, it might just be a story | 00:37 |
| rm_work | so we *could* make the change to install QoS support in our gates... | 00:37 |
| xgerman_ | ok, so we should test QoS… | 00:37 |
| rm_work | hey, we don't TEST it | 00:37 |
| rm_work | but we do install it :P | 00:37 |
| xgerman_ | lol | 00:37 |
| rm_work | we don't have scenarios for Barbican AFAIK | 00:37 |
| rm_work | IIRC | 00:37 |
| rm_work | BBQ | 00:37 |
| johnsom | I think that is the case | 00:37 |
| rm_work | but we DO plan to add some | 00:38 |
| rm_work | at least, I do, and I talked to someone else who does as well | 00:38 |
| johnsom | It has a functional test, but we don't test that neutron actually enables QoS | 00:38 |
| xgerman_ | https://storyboard.openstack.org/#!/story/2001402 | 00:39 |
| xgerman_ | https://storyboard.openstack.org/#!/story/2001401 | 00:40 |
| bar_ | johnsom, xgerman_, I'm still testing QoS. Can you wait with the workflow? | 00:46 |
| xgerman_ | No intention from me to workflow | 00:47 |
| johnsom | Sure, let me know | 00:48 |
| bar_ | johnsom, thx. btw, check out my client-side patch for qos, if you have the time. | 00:53 |
| johnsom | Ah, good point, I still have the devstack | 00:54 |
| *** SumitNaiksatam has quit IRC | 00:56 | |
| rm_work | augh really | 01:07 |
| rm_work | my patch timed out | 01:08 |
| rm_work | on ... dsvm-api | 01:08 |
| rm_work | rechecking and now i get to wait like three hours for a merge T_T | 01:08 |
| bar_ | rm_work, I already rechecked it | 01:11 |
| rm_work | ah lol | 01:11 |
| rm_work | ah well | 01:11 |
| rm_work | i don't think mine will trigger | 01:12 |
| bzhao | good lively channel. :) | 01:12 |
| rm_work | yeah, I like that our channel has ... actual discussions and people :P | 01:13 |
| bzhao | UDP multi keepalived process in a single namespace. T_T | 01:13 |
| rm_work | I still don't understand how it is a problem | 01:14 |
| bzhao | yeah, I like it | 01:14 |
| rm_work | i didn't think it had anything to do with the application itself | 01:14 |
| rm_work | like, i could throw netcat into a namespace too, i thought the app was pretty much unaware | 01:14 |
| bzhao | Yeah, but I remember johonsom said when I intro a new backend, he let me to care about the keepalived sync port. For now, I have no idea about it. I will show you. wair a moment. | 01:15 |
| bzhao | oh s/wair/wait | 01:16 |
| rm_work | yeah but the sync port will be in the namespace if the keepalived process is in the namespace | 01:16 |
| rm_work | but ... | 01:16 |
| rm_work | i don't know if we care | 01:16 |
| rm_work | hmmm, the one thing I would be concerned about in ACTIVE/STANDBY is that keepalived procs need to stay in sync | 01:17 |
| rm_work | because one can't try to move the VIP | 01:17 |
| johnsom | Yeah, I didn't get to looking at that, sigh. The sync is only for the VRRP part | 01:17 |
| rm_work | while the other thinks it is still the master | 01:17 |
| rm_work | right | 01:17 |
| rm_work | and only the *main* keepalived (not in namespace?) would be doing vrrp | 01:17 |
| bzhao | Oh, sorry. maybe I miss something. | 01:17 |
| rm_work | the other ones actually don't even do listening? | 01:17 |
| rm_work | they are for health checks? | 01:18 |
| rm_work | so the entire process should be able to easily run inside the namespace | 01:18 |
| rm_work | with no issues AFAIK | 01:18 |
| rm_work | *ADAIU | 01:18 |
| johnsom | bzhao Well, I confused myself and then probably confused you in the process | 01:18 |
| rm_work | *AFAIU | 01:18 |
| bzhao | johnsom, aha, no matter. :) | 01:18 |
| johnsom | If I remember, there were two things we were looking at. | 01:19 |
| johnsom | 1. Can we run each UDP listener as it's own process so that reload/restarts don't interrupt traffic on the others. | 01:19 |
| johnsom | 2. We should add health check scripts that monitor the UDP keepaliveds as part of the VRRP monitoring. We have a directory for those scripts already, so should be easy. | 01:20 |
| bzhao | http://eavesdrop.openstack.org/irclogs/%23openstack-lbaas/%23openstack-lbaas.2017-12-08.log.html This is my left in last Friday, I'm afraid not fresh the screen. :) | 01:21 |
| rm_work | i think keepalived doesn't have anything to do with the actual listening? since keepalived is just used to configure LVS which does the actual loadbalancing? | 01:22 |
| rm_work | keepalived does the monitoring only | 01:22 |
| xgerman_ | yep, configures LVS and monitors | 01:23 |
| johnsom | Right | 01:23 |
| rm_work | so we can restart keepalived all day | 01:23 |
| bzhao | johnsom, For 1, I think it works, but only for the version 1.3.9 in my test. For 2, I'm working on that, but hang on the current keepalived version I have no idea to run multi process. | 01:23 |
| xgerman_ | BUT keepalived will also check the backend and take it out of rotation | 01:23 |
| rm_work | and it won't stop listening | 01:23 |
| rm_work | or interrupt the members | 01:23 |
| xgerman_ | aka kepplaived does the healthmonitoring of the backend servers | 01:23 |
| johnsom | Yeah, that is probably true... | 01:23 |
| rm_work | yeah so ... honestly it COULD just be one keepalive | 01:23 |
| johnsom | Just not sure 100%. | 01:23 |
| rm_work | well | 01:24 |
| rm_work | two | 01:24 |
| rm_work | one for VRRP and one for inside the namespace for monitoring | 01:24 |
| rm_work | i don't know if you can have one app both inside and not | 01:24 |
| xgerman_ | probably not — keepalived is siple software + you only have one LVS | 01:24 |
| rm_work | yeah | 01:24 |
| rm_work | so LVS needs to be running *inside* the NS, and we need one keepalived outside for VRRP and one inside for healthmonitoring | 01:25 |
| rm_work | it CAN be multiple inside for healthmonitoring (one each) and that is ... fine i think | 01:25 |
| rm_work | i don't know why that would cause problems really | 01:25 |
| rm_work | and it fits our model best | 01:25 |
| xgerman_ | LVS runs in the kernel — so not sure how to make tat run inside the NS | 01:25 |
| bar_ | johnsom, would you mind taking a look at my review: https://review.openstack.org/#/c/458308/ | 01:26 |
| bzhao | rm_work, Currently we implement as listener keepalived 1:1. | 01:26 |
| rm_work | yeah which is easiest | 01:26 |
| rm_work | because it's literally the exact same as if we were doing an haproxy confog | 01:26 |
| rm_work | *config | 01:26 |
| xgerman_ | +1 | 01:26 |
| rm_work | just use a different template for the service config and the systemd config | 01:27 |
| rm_work | and suddenly it should "just work" | 01:27 |
| bzhao | rm_work, yeah. So it is possible to reload/update the config to get all lisnter config should be refresh. | 01:27 |
| johnsom | bar_ still looking at client patch | 01:27 |
| bar_ | xgerman_, doesn't https://storyboard.openstack.org/#!/story/2001310 overlaps with your story? (specifically, task 5863) | 01:30 |
| bzhao | xgerman_, I test in a very sample namesapce env, it works. :). But for higher version keepalived. | 01:30 |
| rm_work | i still don't understand how it wouldn't work for any version | 01:30 |
| rm_work | what are you doing to put it in the namespace? | 01:30 |
| *** sshank has joined #openstack-lbaas | 01:30 | |
| rm_work | are you using a keepalived arg or something? | 01:30 |
| bzhao | rm_work, just some LVS config. | 01:30 |
| rm_work | ... to put keepalived in the namespace? | 01:31 |
| xgerman_ | bzhao: mine is geared toward the gate | 01:31 |
| bzhao | I just test with very simple keepalived config | 01:31 |
| johnsom | bar_ how do I un-set the QoS policy with the client? | 01:31 |
| xgerman_ | for instance we enable barbican in the gate but not in the devstack somone might install on ntheir box | 01:31 |
| bar_ | johnsom, You cannot | 01:31 |
| bzhao | This is my new keepalived config. http://paste.openstack.org/show/628411/ | 01:32 |
| bzhao | xgerman_, sorry. :) | 01:32 |
| bzhao | This is the existing keepalived config. It was still running | 01:32 |
| bzhao | http://paste.openstack.org/show/628412/ | 01:32 |
| bar_ | johnsom, nor the server side would allow that at this point. | 01:32 |
| xgerman_ | no worries | 01:32 |
| bzhao | I run keepalived like "ip netns exec amphora-haproxy keepalived -D -d -f config". | 01:32 |
| bzhao | Then I saw the log in syslog like: http://paste.openstack.org/show/628413/ | 01:32 |
| johnsom | bar_ Yes, I can un-set it by passing "null" | 01:33 |
| rm_work | yes | 01:33 |
| bzhao | Also, the higher version of keepalived, such as 1.3.9 It supports "net_namespace" and "instance" configuration options. Fo runing multiple keepalveds in the same namespace. | 01:33 |
| rm_work | bzhao: because you are trying to use a central daemon and you already have one | 01:33 |
| rm_work | i don't know how it would work with any version of keepalived using this method | 01:33 |
| bar_ | johnsom, ha. I missed that. | 01:33 |
| rm_work | you need to differentiate them | 01:33 |
| johnsom | bar_ I think we are missing a few unsets in other spots too | 01:34 |
| bzhao | rm_work, eh, you mean I need to kill the exist first and recreate them? | 01:34 |
| rm_work | no | 01:34 |
| bar_ | johnsom, what do you mean? | 01:34 |
| johnsom | bar_ I think other commands should probably have unset. Not a worry on this patch though | 01:34 |
| rm_work | bzhao: try for example running it with `-n` in a screen | 01:35 |
| rm_work | you should then be able to run as many as you want | 01:35 |
| rm_work | `ip netns exec amphora-haproxy keepalived -n -D -d -f config` | 01:35 |
| rm_work | though why do you have '-d'? | 01:36 |
| bar_ | johnsom, If you think we're missing a valuable feature (like unset), please open a story and make sure to ping\mail me about that. | 01:36 |
| rm_work | bzhao: it won't fork a daemon, but that's fine, when we're running it for real we'll take care of that in the servicefile | 01:36 |
| bzhao | rm_work, aha, There maybe some mistakes I make, I saw the existing code to run a keepalived process with 'ip netns exec keepalived -D -d', so I thought it could run keepalived in a ns..T_T | 01:38 |
| rm_work | it can | 01:38 |
| rm_work | but running more than ONE (in a namespace or not) means you can't use it as a daemon quite the same way | 01:38 |
| rm_work | because it's trying to control a single instance | 01:38 |
| rm_work | possibly actually you could do this | 01:39 |
| bzhao | rm_work, Thanks for cliear. | 01:39 |
| rm_work | `ip netns exec amphora-haproxy keepalived -D -d -f config1 -p /var/run/keepalive1.pid` | 01:40 |
| rm_work | `ip netns exec amphora-haproxy keepalived -D -d -f config2 -p /var/run/keepalive2.pid` | 01:40 |
| rm_work | that might work? i'm not sure | 01:40 |
| rm_work | because by default it uses `/var/run/keepalived.pid` and that means it'll try to use the same instance | 01:40 |
| rm_work | again, when we make a service file for it, it'll take care of this, so it's really a non-issue | 01:41 |
| bzhao | I just test with it if there is already a -D keepalived process, it won't work | 01:41 |
| *** yamamoto has joined #openstack-lbaas | 01:41 | |
| johnsom | bzhao I think it is the pid file that is giving you the "already running" errors | 01:41 |
| rm_work | bzhao: if you specify `-p mypidfile` it should work | 01:42 |
| bzhao | the pid file not exist, as the process not trully setup | 01:42 |
| rm_work | possibly you also need `-r` and `-c` | 01:42 |
| rm_work | it should create the pid file | 01:42 |
| rm_work | sec, let me test this really quick | 01:42 |
| johnsom | My question is though, since this is configuring lvs, we might require a special version of the kernel with lvs namespace support and/or the keepalived that is namespace aware... | 01:43 |
| rm_work | hmmm | 01:43 |
| rm_work | i think we need to have LVS start in the namespace | 01:43 |
| rm_work | but yeah that's interesting | 01:43 |
| rm_work | so FYI the xenial version of keepalived is namespace aware... | 01:43 |
| johnsom | https://lwn.net/Articles/419692/ | 01:43 |
| rm_work | that was in 2010 | 01:44 |
| rm_work | if that isn't merged, I'll cry | 01:44 |
| johnsom | Oh, I know the stuff is there, I just don't know what magic incantation of versions we need | 01:44 |
| bzhao | Yeah, so just want to raise it. As I know if we update the keepalived version maybe risky. | 01:45 |
| rm_work | bzhao: yeah i just tested it and it works | 01:50 |
| bar_ | johnsom, can you take a look at https://review.openstack.org/#/c/458308/ now? I'm not sure $2 in my review is actually a bug or not. | 01:51 |
| rm_work | http://paste.openstack.org/show/628665/ | 01:51 |
| rm_work | bzhao: ^^ | 01:51 |
| bzhao | rm_work, cool, whay options follow.. | 01:51 |
| rm_work | that will work in a namespace or not | 01:52 |
| rm_work | just add the namespace code before it | 01:52 |
| rm_work | it's that they can't share a pidfile and there is a default file it will look for | 01:52 |
| johnsom | bar_ I am looking at it now, but I'm struggling to see how the code behaves that way | 01:52 |
| bzhao | rm_work, oh, i know, I should split the check subprocess pid either. | 01:53 |
| bzhao | rm_work, I jsut test -p and -r. | 01:54 |
| bzhao | rm_work, it won't work. | 01:54 |
| *** sshank has quit IRC | 01:56 | |
| rm_work | well, i know this works, because it is running on my machine this way :) | 01:56 |
| bzhao | rm_work, then there is 1 thing need to clear. If we want to implement udp for 1 listener : 1 keepalived process. aha :) | 01:57 |
| *** Swami has quit IRC | 01:58 | |
| rm_work | i think it should be fine | 01:58 |
| bar_ | johnsom, I would expect it to apply the qos policy specifically on the vip_port inside the ApplyQos network task, which it doesn't currently | 01:58 |
| johnsom | bar_ Oh! Are you looking at the VIP port or the base port? I bet that is the issue. Test the qos through the load balancer, I bet it is limited by what was set in the openstack loadbalnacer call | 02:00 |
| johnsom | QoS doesn't really apply to AAP ports, they are just fake ports to track an IP. The QoS has to be applied to the base port. | 02:00 |
| bzhao | OK, thanks johnsom and rm_work for many many patience to solve my problem. I will finish the agent part coding this week with some test, i think. | 02:01 |
| johnsom | bar_ you will notice the "VIP" AAP port is always "down" | 02:01 |
| bzhao | Yeah, it just apply qos on the vrrp_port on amp instances | 02:02 |
| bzhao | The vip port not bound in neutron. We just use it for allowed address pair for VIP function | 02:02 |
| bar_ | johnsom, ok, thanks for looking into it. Not-a-bug then. | 02:03 |
| bzhao | You guys message are so quick, T_T.. the qos patch, also metioned by german, if possible I can take over the work for doc and CI. | 02:05 |
| *** reedip has quit IRC | 02:06 | |
| *** bar_ has quit IRC | 02:06 | |
| *** annp has joined #openstack-lbaas | 02:15 | |
| *** reedip has joined #openstack-lbaas | 02:17 | |
| bzhao | johnsom, xgerman_, Should I post a quick new patch for Qos patch? or waiting for your kind test? | 02:18 |
| johnsom | bzhao go for it | 02:20 |
| openstackgerrit | ZhaoBo proposed openstack/octavia master: Extend api to accept qos_policy_id https://review.openstack.org/458308 | 02:25 |
| rm_work | aug gonna need to recheck again >_< | 02:30 |
| rm_work | gate got a bad host or something i think | 02:30 |
| rm_work | http://logs.openstack.org/90/525790/6/gate/octavia-v1-dsvm-scenario/96ecd28/logs/screen-n-cpu.txt.gz#_Dec_12_02_08_21_221163 | 02:31 |
| rm_work | wtf | 02:31 |
| *** sanfern has quit IRC | 02:39 | |
| *** dmellado has quit IRC | 02:42 | |
| *** tongl has joined #openstack-lbaas | 02:43 | |
| *** SumitNaiksatam has joined #openstack-lbaas | 02:44 | |
| *** dmellado has joined #openstack-lbaas | 02:48 | |
| *** bbbbzhao_ has joined #openstack-lbaas | 02:52 | |
| openstackgerrit | Hengqing Hu proposed openstack/octavia-dashboard master: Remove members with a pool id https://review.openstack.org/527286 | 02:58 |
| openstackgerrit | hujin proposed openstack/neutron-lbaas master: Remove key "l7_policies" in pool dict https://review.openstack.org/527287 | 02:58 |
| openstackgerrit | hujin proposed openstack/neutron-lbaas master: Remove key "l7_policies" in pool dict https://review.openstack.org/527287 | 03:00 |
| *** ianychoi has joined #openstack-lbaas | 03:08 | |
| *** sanfern has joined #openstack-lbaas | 03:27 | |
| *** dmellado has quit IRC | 03:37 | |
| *** dmellado has joined #openstack-lbaas | 03:38 | |
| sanfern | hi johnsom | 03:47 |
| *** armax has quit IRC | 03:53 | |
| *** armax has joined #openstack-lbaas | 03:53 | |
| *** armax has quit IRC | 03:54 | |
| *** krypto has joined #openstack-lbaas | 04:51 | |
| *** robcresswell has quit IRC | 05:39 | |
| *** krypto has quit IRC | 05:39 | |
| *** krypto has joined #openstack-lbaas | 05:40 | |
| *** krypto has quit IRC | 05:40 | |
| *** krypto has joined #openstack-lbaas | 05:40 | |
| *** numans has quit IRC | 05:54 | |
| *** numans has joined #openstack-lbaas | 05:55 | |
| openstackgerrit | Hengqing Hu proposed openstack/octavia master: Extend api to accept qos_policy_id https://review.openstack.org/458308 | 05:59 |
| *** krypto has quit IRC | 06:05 | |
| *** SumitNaiksatam has quit IRC | 06:12 | |
| *** tongl has quit IRC | 06:13 | |
| *** krypto has joined #openstack-lbaas | 06:22 | |
| *** krypto has quit IRC | 06:22 | |
| *** krypto has joined #openstack-lbaas | 06:22 | |
| *** krypto has quit IRC | 06:28 | |
| *** krypto has joined #openstack-lbaas | 06:29 | |
| *** krypto has quit IRC | 06:29 | |
| *** krypto has joined #openstack-lbaas | 06:29 | |
| *** bbbbzhao_ has quit IRC | 06:42 | |
| *** sanfern has quit IRC | 06:59 | |
| *** sanfern has joined #openstack-lbaas | 06:59 | |
| *** rcernin has quit IRC | 07:21 | |
| *** dokua has joined #openstack-lbaas | 07:31 | |
| *** dokua has quit IRC | 07:36 | |
| *** kobis has joined #openstack-lbaas | 07:39 | |
| *** Alex_Staf has joined #openstack-lbaas | 07:39 | |
| *** robcresswell has joined #openstack-lbaas | 07:39 | |
| *** dokua has joined #openstack-lbaas | 07:44 | |
| *** Alex_Staf has quit IRC | 07:48 | |
| *** AlexeyAbashkin has joined #openstack-lbaas | 07:53 | |
| rm_work | bah is the gate having issues | 08:04 |
| *** b_bezak has joined #openstack-lbaas | 08:06 | |
| *** tesseract has joined #openstack-lbaas | 08:22 | |
| *** Alex_Staf has joined #openstack-lbaas | 08:22 | |
| *** rcernin has joined #openstack-lbaas | 08:36 | |
| *** krypto has quit IRC | 08:43 | |
| -openstackstatus- NOTICE: Our CI system Zuul is currently not accessible. Wait with approving changes and rechecks until it's back online. Currently waiting for an admin to investigate. | 08:48 | |
| *** krypto has joined #openstack-lbaas | 08:59 | |
| *** krypto has quit IRC | 08:59 | |
| *** krypto has joined #openstack-lbaas | 08:59 | |
| *** Alex_Staf has quit IRC | 09:05 | |
| -openstackstatus- NOTICE: Zuul is back online, looks like a temporary network problem. | 09:08 | |
| *** jappleii__ has quit IRC | 09:10 | |
| *** dmellado has quit IRC | 09:17 | |
| *** strigazi_ is now known as strigazi | 09:17 | |
| *** krypto has quit IRC | 09:17 | |
| *** krypto has joined #openstack-lbaas | 09:17 | |
| *** krypto has quit IRC | 09:17 | |
| *** krypto has joined #openstack-lbaas | 09:17 | |
| *** devfaz has quit IRC | 09:37 | |
| *** devfaz has joined #openstack-lbaas | 09:38 | |
| openstackgerrit | ZhaoBo proposed openstack/octavia-tempest-plugin master: [WIP]new CI for test tempest with neutron and octavia https://review.openstack.org/527356 | 09:41 |
| *** dmellado has joined #openstack-lbaas | 09:48 | |
| *** eN_Guruprasad_Rn has joined #openstack-lbaas | 09:54 | |
| *** annp has quit IRC | 09:56 | |
| *** dmellado has quit IRC | 10:07 | |
| *** dmellado has joined #openstack-lbaas | 10:20 | |
| *** sanfern has quit IRC | 10:20 | |
| *** salmankhan has joined #openstack-lbaas | 10:24 | |
| *** yamamoto has quit IRC | 10:35 | |
| *** Alex_Staf has joined #openstack-lbaas | 10:35 | |
| rm_work | looks like zuul still f'd | 10:37 |
| rm_work | when it comes back fully we can recheck stuff i guess T_T | 10:38 |
| *** yamamoto has joined #openstack-lbaas | 10:48 | |
| *** dmellado has quit IRC | 10:50 | |
| *** Alex_Staf has quit IRC | 10:52 | |
| *** sticker has joined #openstack-lbaas | 10:59 | |
| *** dmellado has joined #openstack-lbaas | 10:59 | |
| bzhao | Yeah, it dead all day | 11:08 |
| *** AlexeyAbashkin has quit IRC | 11:15 | |
| *** dmellado has quit IRC | 11:16 | |
| *** dmellado has joined #openstack-lbaas | 11:18 | |
| *** dmellado has quit IRC | 11:20 | |
| *** dmellado has joined #openstack-lbaas | 11:24 | |
| nmagnezi | rm_work, o/ | 11:28 |
| openstackgerrit | Santhosh Fernandes proposed openstack/octavia master: [WIP] L3 ACTIVE-ACTIVE Data model impact https://review.openstack.org/524722 | 11:29 |
| *** dmellado has quit IRC | 11:32 | |
| *** dmellado has joined #openstack-lbaas | 11:33 | |
| *** salmankhan has quit IRC | 11:34 | |
| *** AlexeyAbashkin has joined #openstack-lbaas | 11:40 | |
| *** salmankhan has joined #openstack-lbaas | 11:43 | |
| *** yamamoto has quit IRC | 11:54 | |
| *** yamamoto has joined #openstack-lbaas | 11:58 | |
| *** yamamoto has quit IRC | 12:03 | |
| nmagnezi | rm_work, around? | 12:17 |
| *** sticker has quit IRC | 12:34 | |
| *** yamamoto has joined #openstack-lbaas | 12:35 | |
| *** atoth has joined #openstack-lbaas | 13:22 | |
| *** dokua has quit IRC | 14:02 | |
| *** b3nt_pin is now known as beagles | 14:07 | |
| *** yamamoto has quit IRC | 14:24 | |
| *** yamamoto has joined #openstack-lbaas | 14:24 | |
| *** salmankhan has quit IRC | 14:25 | |
| *** tesseract has quit IRC | 14:26 | |
| *** gcheresh has joined #openstack-lbaas | 14:26 | |
| *** eN_Guruprasad_Rn has quit IRC | 14:27 | |
| *** salmankhan has joined #openstack-lbaas | 14:38 | |
| -openstackstatus- NOTICE: We're currently seeing an elevated rate of timeouts in jobs and the zuulv3.openstack.org dashboard is intermittently unresponsive, please stand by while we troubleshoot the issues. | 14:39 | |
| *** krypto has quit IRC | 14:49 | |
| *** salmankhan has quit IRC | 14:53 | |
| *** alee has quit IRC | 14:55 | |
| *** salmankhan has joined #openstack-lbaas | 15:00 | |
| *** rcernin has quit IRC | 15:00 | |
| *** sanfern has joined #openstack-lbaas | 15:01 | |
| *** dokua has joined #openstack-lbaas | 15:37 | |
| *** salmankhan has quit IRC | 15:37 | |
| *** gcheresh has quit IRC | 15:45 | |
| *** tongl has joined #openstack-lbaas | 15:47 | |
| *** dokua has quit IRC | 15:52 | |
| *** dokua has joined #openstack-lbaas | 15:53 | |
| *** kobis has quit IRC | 15:53 | |
| *** salmankhan has joined #openstack-lbaas | 15:55 | |
| *** tongl has quit IRC | 15:55 | |
| *** dokua has quit IRC | 15:57 | |
| *** fnaval has joined #openstack-lbaas | 16:03 | |
| *** b_bezak has quit IRC | 16:04 | |
| *** b_bezak has joined #openstack-lbaas | 16:04 | |
| *** b_bezak has quit IRC | 16:09 | |
| *** armax has joined #openstack-lbaas | 16:26 | |
| *** AlexeyAbashkin has quit IRC | 16:26 | |
| *** AlexeyAbashkin has joined #openstack-lbaas | 16:38 | |
| *** sanfern has quit IRC | 16:43 | |
| *** sanfern has joined #openstack-lbaas | 16:44 | |
| *** AlexeyAbashkin has quit IRC | 16:47 | |
| sanfern | hi johnsom | 16:50 |
| johnsom | Hi Sanfern | 16:51 |
| sanfern | I saw your review comments , thank you for comments. I need your help to resolve these. | 16:51 |
| sanfern | https://review.openstack.org/#/c/490164/21/elements/exabgp-speaker/post-install.d/77-exabgp - I tried removing pip install statement and rebuild the image. VM didn't install exabgp | 16:53 |
| sanfern | I am making mkfifo -m 0644 it should be fine right | 16:55 |
| sanfern | should be render exabgp.service file and remove it from the element ? It doesn't start by default | 16:57 |
| johnsom | Hmmm, ok, so if the exabgp service isn't enabled, that is fine, you can leave it. We just need to manage enabling it on the agent | 17:06 |
| sanfern | ok I am disabling when I stop the service | 17:07 |
| johnsom | The pip install thing, that doesn't make sense as it's in the octavia requirements.txt. The only thing I can think is it's installing in py35 and you are looking in py27 or vs versa | 17:07 |
| johnsom | Yeah, but I never saw it enabled | 17:07 |
| johnsom | You might want to go tighter than 644, as it's a fifo, so if someone else reads from it you will not see that data. Why not 600? | 17:08 |
| sanfern | ok I have not tested on py35. I need to verify it. | 17:08 |
| johnsom | I can look at an image later and see what is up. I just can't right now. | 17:09 |
| sanfern | I will retest it. tomorrow in office. | 17:09 |
| sanfern | https://review.openstack.org/#/c/491016/13/octavia/amphorae/backends/agent/api_server/exabgp.py@36 | 17:09 |
| johnsom | Ok. Maybe I can get there this afternoon. | 17:10 |
| sanfern | I need to make entry in interface file | 17:10 |
| *** tongl has joined #openstack-lbaas | 17:10 | |
| *** salmankhan has quit IRC | 17:31 | |
| *** salmankhan has joined #openstack-lbaas | 17:37 | |
| *** Dave has quit IRC | 17:37 | |
| *** dokua has joined #openstack-lbaas | 17:38 | |
| *** dokua has quit IRC | 17:47 | |
| openstackgerrit | Santhosh Fernandes proposed openstack/octavia master: [WIP] L3 ACTIVE-ACTIVE Data model impact https://review.openstack.org/524722 | 17:49 |
| *** Dave has joined #openstack-lbaas | 17:53 | |
| sanfern | johnsom, http://paste.openstack.org/show/g82NMaPHGF1YA3iHxUbQ/ any clue to fix it ? src - https://review.openstack.org/#/c/524722/13/octavia/tests/functional/db/test_repositories.py | 17:53 |
| *** fnaval has quit IRC | 17:53 | |
| sanfern | in deleting Distributor flow, should we remove from the DB or mark for deletion ? | 17:54 |
| johnsom | That error is interesting, first glance I don't see the issue. | 18:01 |
| johnsom | Mark for deletion is best, it gives an operator a path back if something goes wrong | 18:01 |
| sanfern | thanks johnsom | 18:02 |
| sanfern | my all repo test on that patch is failing :( | 18:02 |
| openstackgerrit | Santhosh Fernandes proposed openstack/octavia master: Adding exabgp-speaker element to amphora image https://review.openstack.org/490164 | 18:07 |
| *** AlexeyAbashkin has joined #openstack-lbaas | 18:14 | |
| *** AlexeyAbashkin has quit IRC | 18:18 | |
| *** gcheresh has joined #openstack-lbaas | 18:33 | |
| *** AlexeyAbashkin has joined #openstack-lbaas | 18:34 | |
| *** AlexeyAbashkin has quit IRC | 18:38 | |
| *** dokua has joined #openstack-lbaas | 18:39 | |
| rm_work | nmagnezi: i am now... | 18:40 |
| *** Swami has joined #openstack-lbaas | 18:41 | |
| *** dokua has quit IRC | 18:42 | |
| *** dokua has joined #openstack-lbaas | 18:42 | |
| *** dokua has quit IRC | 18:42 | |
| rm_work | bleh zuul still f'd | 18:50 |
| *** salmankhan has quit IRC | 19:34 | |
| *** dokua has joined #openstack-lbaas | 19:47 | |
| *** Alex_Staf has joined #openstack-lbaas | 19:47 | |
| *** Alex_Staf has quit IRC | 19:53 | |
| *** dokua has quit IRC | 20:05 | |
| -openstackstatus- NOTICE: The zuul scheduler has been restarted after lengthy troubleshooting for a memory consumption issue; earlier changes have been reenqueued but if you notice jobs not running for a new or approved change you may want to leave a recheck comment or a new approval vote | 20:15 | |
| *** dmellado has quit IRC | 20:32 | |
| *** openstackgerrit has quit IRC | 20:34 | |
| *** dmellado has joined #openstack-lbaas | 20:34 | |
| *** sbalukoff_ has quit IRC | 20:34 | |
| *** jmccrory has quit IRC | 20:35 | |
| *** sbalukoff_ has joined #openstack-lbaas | 20:37 | |
| *** jmccrory has joined #openstack-lbaas | 20:40 | |
| *** dokua has joined #openstack-lbaas | 21:00 | |
| *** jappleii__ has joined #openstack-lbaas | 21:23 | |
| *** jappleii__ has quit IRC | 21:24 | |
| *** jappleii__ has joined #openstack-lbaas | 21:25 | |
| *** jappleii__ has quit IRC | 21:26 | |
| *** jappleii__ has joined #openstack-lbaas | 21:27 | |
| *** jappleii__ has quit IRC | 21:27 | |
| *** jappleii__ has joined #openstack-lbaas | 21:28 | |
| *** AlexeyAbashkin has joined #openstack-lbaas | 21:34 | |
| *** AlexeyAbashkin has quit IRC | 21:38 | |
| *** gcheresh has quit IRC | 21:41 | |
| rm_work | anyone actually here? am I netsplit? :P | 21:52 |
| xgerman_ | Na | 21:53 |
| dims | rm_work o/ | 21:54 |
| rm_work | k :P thanks | 21:54 |
| rm_work | unnaturally quiet | 21:54 |
| *** dokua has quit IRC | 21:56 | |
| *** sticker has joined #openstack-lbaas | 21:57 | |
| *** AlexeyAbashkin has joined #openstack-lbaas | 21:57 | |
| *** rcernin has joined #openstack-lbaas | 21:58 | |
| *** AlexeyAbashkin has quit IRC | 22:02 | |
| johnsom | True, I got pulled into an internal issue this morning, so just now getting back to the day job. | 22:02 |
| johnsom | grin | 22:02 |
| *** sticker has quit IRC | 22:10 | |
| *** sticker has joined #openstack-lbaas | 22:10 | |
| nmagnezi | johnsom, just replied to the providers spec | 22:18 |
| johnsom | Thanks | 22:18 |
| nmagnezi | johnsom, i think it's mostly done. great job there. just few minor things IMO | 22:18 |
| nmagnezi | np at all :) | 22:18 |
| nmagnezi | rm_work, someday.. we'll manage to chat here :-) | 22:21 |
| rm_work | lol | 22:23 |
| rm_work | nmagnezi: how about now? going to bed? :P | 22:24 |
| nmagnezi | rm_work, yeah.. sorry :-) | 22:24 |
| rm_work | np heh | 22:24 |
| *** openstackgerrit has joined #openstack-lbaas | 22:46 | |
| openstackgerrit | Merged openstack/octavia master: Move loading the network driver into the flows https://review.openstack.org/525790 | 22:46 |
| rm_work | yay finally, k | 22:47 |
| rm_work | aaaand qos is in conflict :( | 22:47 |
| rm_work | sorry about that | 22:47 |
| rm_work | that's part of why i was hoping that would merge quickly, and these delays were killing me | 22:48 |
| openstackgerrit | Adam Harwell proposed openstack/octavia master: Producer/endpoint code to allow for amphora failovers https://review.openstack.org/525302 | 22:48 |
| openstackgerrit | Adam Harwell proposed openstack/octavia master: Amphora API Failover call https://review.openstack.org/525778 | 22:48 |
| openstackgerrit | Adam Harwell proposed openstack/octavia master: Add unit tests for neutron utils, add model/util for floating_ip https://review.openstack.org/525353 | 22:48 |
| rm_work | rebasinggggg | 22:49 |
| *** fnaval has joined #openstack-lbaas | 23:06 | |
| openstackgerrit | Adam Harwell proposed openstack/octavia master: WIP: Floating IP Network Driver (spans L3s) https://review.openstack.org/435612 | 23:15 |
| *** chandankumar has quit IRC | 23:22 | |
| *** chandankumar has joined #openstack-lbaas | 23:24 | |
| *** chandankumar has quit IRC | 23:29 | |
| *** chandankumar has joined #openstack-lbaas | 23:35 | |
| *** atoth has quit IRC | 23:35 | |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!