| *** tetsuro has joined #openstack-meeting | 00:21 | |
| *** maohongbo1 has joined #openstack-meeting | 00:48 | |
| *** maohongbo has quit IRC | 00:50 | |
| *** maohongbo1 is now known as maohongbo | 00:50 | |
| *** ociuhandu has joined #openstack-meeting | 01:01 | |
| *** jamesmcarthur has joined #openstack-meeting | 01:01 | |
| *** ysandeep|away is now known as ysandeep | 01:03 | |
| *** maohongbo1 has joined #openstack-meeting | 01:04 | |
| *** ociuhandu has quit IRC | 01:05 | |
| *** maohongbo has quit IRC | 01:06 | |
| *** maohongbo1 is now known as maohongbo | 01:06 | |
| *** Liang__ has joined #openstack-meeting | 01:10 | |
| *** ricolin has joined #openstack-meeting | 01:14 | |
| *** jamesmcarthur has quit IRC | 01:24 | |
| *** maohongbo1 has joined #openstack-meeting | 01:28 | |
| *** maohongbo has quit IRC | 01:30 | |
| *** maohongbo1 is now known as maohongbo | 01:30 | |
| *** maohongbo1 has joined #openstack-meeting | 01:36 | |
| *** maohongbo has quit IRC | 01:38 | |
| *** maohongbo1 is now known as maohongbo | 01:38 | |
| *** yaawang has quit IRC | 01:46 | |
| *** yaawang has joined #openstack-meeting | 01:47 | |
| *** yaawang has quit IRC | 01:52 | |
| *** yaawang has joined #openstack-meeting | 01:53 | |
| *** rbudden has quit IRC | 02:05 | |
| *** tetsuro has quit IRC | 02:06 | |
| *** maohongbo has quit IRC | 02:08 | |
| *** apetrich has quit IRC | 02:08 | |
| *** maohongbo has joined #openstack-meeting | 02:08 | |
| *** tetsuro has joined #openstack-meeting | 02:12 | |
| *** jmasud has joined #openstack-meeting | 02:15 | |
| *** jmasud has quit IRC | 02:20 | |
| *** ysandeep is now known as ysandeep|afk | 02:21 | |
| *** jmasud has joined #openstack-meeting | 02:36 | |
| *** igordc has joined #openstack-meeting | 02:46 | |
| *** jmasud has quit IRC | 02:50 | |
| *** igordc has quit IRC | 02:52 | |
| *** igordc has joined #openstack-meeting | 02:52 | |
| *** igordc has quit IRC | 02:57 | |
| *** jmasud has joined #openstack-meeting | 03:00 | |
| *** ykatabam has quit IRC | 03:10 | |
| *** rcernin has quit IRC | 03:10 | |
| *** psachin has joined #openstack-meeting | 03:17 | |
| *** jamesmcarthur has joined #openstack-meeting | 03:24 | |
| *** rcernin has joined #openstack-meeting | 03:25 | |
| *** ttsiouts_ has joined #openstack-meeting | 03:31 | |
| *** ttsiouts has quit IRC | 03:34 | |
| *** andrebeltrami has quit IRC | 03:39 | |
| *** ykatabam has joined #openstack-meeting | 03:39 | |
| *** yaawang has quit IRC | 03:40 | |
| *** yaawang has joined #openstack-meeting | 03:41 | |
| *** ociuhandu has joined #openstack-meeting | 03:44 | |
| *** rh-jelabarre has quit IRC | 03:48 | |
| *** ociuhandu has quit IRC | 03:49 | |
| *** tetsuro has quit IRC | 03:51 | |
| *** jmasud has quit IRC | 04:02 | |
| *** jmasud has joined #openstack-meeting | 04:14 | |
| *** ysandeep|afk is now known as ysandeep | 04:22 | |
| *** kevinz has joined #openstack-meeting | 04:42 | |
| *** ysandeep is now known as ysandeep|reboot | 04:49 | |
| *** ysandeep|reboot is now known as ysandeep | 04:53 | |
| *** jmasud has quit IRC | 04:55 | |
| *** jmasud has joined #openstack-meeting | 04:56 | |
| *** vishakha has quit IRC | 05:10 | |
| *** mnasiadka has quit IRC | 05:10 | |
| *** thgcorrea has quit IRC | 05:10 | |
| *** tinwood has quit IRC | 05:10 | |
| *** tetsuro has joined #openstack-meeting | 05:13 | |
| *** vishakha has joined #openstack-meeting | 05:15 | |
| *** mnasiadka has joined #openstack-meeting | 05:15 | |
| *** thgcorrea has joined #openstack-meeting | 05:15 | |
| *** tinwood has joined #openstack-meeting | 05:15 | |
| *** jamesmcarthur has quit IRC | 05:30 | |
| *** jamesmcarthur has joined #openstack-meeting | 05:30 | |
| *** ociuhandu has joined #openstack-meeting | 05:37 | |
| *** ociuhandu has quit IRC | 05:47 | |
| *** ysandeep is now known as ysandeep|brb | 05:49 | |
| *** jamesmcarthur has quit IRC | 05:53 | |
| *** gyee has quit IRC | 05:53 | |
| *** jmasud has quit IRC | 05:56 | |
| *** jmasud has joined #openstack-meeting | 06:11 | |
| *** ysandeep|brb is now known as ysandeep | 06:12 | |
| *** jmasud has quit IRC | 06:13 | |
| *** jmasud has joined #openstack-meeting | 06:15 | |
| *** dklyle has quit IRC | 06:26 | |
| *** ociuhandu has joined #openstack-meeting | 06:37 | |
| *** jamesmcarthur has joined #openstack-meeting | 06:43 | |
| *** ociuhandu has quit IRC | 06:43 | |
| *** jamesmcarthur has quit IRC | 06:47 | |
| *** maohongbo1 has joined #openstack-meeting | 06:51 | |
| *** maohongbo has quit IRC | 06:54 | |
| *** maohongbo1 is now known as maohongbo | 06:54 | |
| *** jamesmcarthur has joined #openstack-meeting | 06:55 | |
| *** jamesmcarthur has quit IRC | 06:59 | |
| *** links has joined #openstack-meeting | 07:00 | |
| *** jamesmcarthur has joined #openstack-meeting | 07:01 | |
| *** jhesketh has quit IRC | 07:04 | |
| *** jamesmcarthur has quit IRC | 07:05 | |
| *** jamesmcarthur has joined #openstack-meeting | 07:07 | |
| *** ociuhandu has joined #openstack-meeting | 07:09 | |
| *** jamesmcarthur has quit IRC | 07:11 | |
| *** slaweq has joined #openstack-meeting | 07:11 | |
| *** jamesmcarthur has joined #openstack-meeting | 07:13 | |
| *** maohongbo1 has joined #openstack-meeting | 07:14 | |
| *** maohongbo has quit IRC | 07:15 | |
| *** maohongbo1 is now known as maohongbo | 07:15 | |
| *** jamesmcarthur has quit IRC | 07:17 | |
| *** jamesmcarthur has joined #openstack-meeting | 07:19 | |
| *** rpittau|afk is now known as rpittau | 07:19 | |
| *** apetrich has joined #openstack-meeting | 07:21 | |
| *** jamesmcarthur has quit IRC | 07:23 | |
| *** maciejjozefczyk has joined #openstack-meeting | 07:24 | |
| *** ociuhandu has quit IRC | 07:30 | |
| *** ralonsoh has joined #openstack-meeting | 07:38 | |
| *** jamesmcarthur has joined #openstack-meeting | 07:46 | |
| *** jamesmcarthur has quit IRC | 07:46 | |
| *** jamesmcarthur has joined #openstack-meeting | 07:46 | |
| *** jamesmcarthur has quit IRC | 07:49 | |
| *** maohongbo has quit IRC | 07:49 | |
| *** maohongbo1 has joined #openstack-meeting | 07:49 | |
| *** jamesmcarthur has joined #openstack-meeting | 07:49 | |
| *** maohongbo1 is now known as maohongbo | 07:51 | |
| *** jamesmcarthur has quit IRC | 07:57 | |
| *** ysandeep is now known as ysandeep|lunch | 07:57 | |
| *** ociuhandu has joined #openstack-meeting | 08:04 | |
| *** ttsiouts_ has quit IRC | 08:11 | |
| *** e0ne has joined #openstack-meeting | 08:13 | |
| *** ttsiouts has joined #openstack-meeting | 08:21 | |
| *** tetsuro has quit IRC | 08:22 | |
| *** ociuhandu has quit IRC | 08:23 | |
| *** ociuhandu has joined #openstack-meeting | 08:23 | |
| *** ysandeep|lunch is now known as ysandeep | 08:25 | |
| *** ykatabam has quit IRC | 08:27 | |
| *** jamesmcarthur has joined #openstack-meeting | 08:28 | |
| *** e0ne_ has joined #openstack-meeting | 08:30 | |
| *** e0ne has quit IRC | 08:30 | |
| *** jmasud has quit IRC | 08:33 | |
| *** maciejjozefczyk has quit IRC | 08:37 | |
| *** ociuhandu has quit IRC | 08:45 | |
| *** ttsiouts has quit IRC | 08:47 | |
| *** maohongbo1 has joined #openstack-meeting | 08:48 | |
| *** maohongbo has quit IRC | 08:48 | |
| *** maohongbo1 is now known as maohongbo | 08:49 | |
| *** ykatabam has joined #openstack-meeting | 09:00 | |
| *** ttsiouts has joined #openstack-meeting | 09:00 | |
| *** ykatabam has quit IRC | 09:08 | |
| *** maohongbo1 has joined #openstack-meeting | 09:11 | |
| *** maohongbo has quit IRC | 09:12 | |
| *** maohongbo1 is now known as maohongbo | 09:12 | |
| *** ttsiouts has quit IRC | 09:16 | |
| *** ysandeep is now known as ysandeep|afk | 09:21 | |
| *** ociuhandu has joined #openstack-meeting | 09:25 | |
| *** maohongbo has quit IRC | 09:32 | |
| *** maohongbo has joined #openstack-meeting | 09:33 | |
| *** jamesmcarthur has quit IRC | 09:34 | |
| *** ttsiouts has joined #openstack-meeting | 09:39 | |
| *** ociuhandu has quit IRC | 09:43 | |
| *** ociuhandu has joined #openstack-meeting | 09:49 | |
| *** Liang__ has quit IRC | 09:52 | |
| *** jamesmcarthur has joined #openstack-meeting | 10:03 | |
| *** ykatabam has joined #openstack-meeting | 10:05 | |
| *** e0ne_ has quit IRC | 10:21 | |
| *** e0ne has joined #openstack-meeting | 10:22 | |
| *** rpittau is now known as rpittau|bbl | 10:30 | |
| *** brinzhang has quit IRC | 10:35 | |
| *** ociuhandu has quit IRC | 10:48 | |
| *** ociuhandu has joined #openstack-meeting | 10:48 | |
| *** ttsiouts has quit IRC | 10:49 | |
| *** ttsiouts has joined #openstack-meeting | 10:50 | |
| *** ociuhandu has quit IRC | 10:52 | |
| *** ykatabam has quit IRC | 11:09 | |
| *** ykatabam has joined #openstack-meeting | 11:15 | |
| *** bbowen_ has joined #openstack-meeting | 11:18 | |
| *** bbowen has quit IRC | 11:19 | |
| *** ociuhandu has joined #openstack-meeting | 11:23 | |
| *** e0ne has quit IRC | 11:25 | |
| *** e0ne has joined #openstack-meeting | 11:25 | |
| *** bbowen has joined #openstack-meeting | 11:37 | |
| *** bbowen_ has quit IRC | 11:39 | |
| *** rfolco has joined #openstack-meeting | 11:47 | |
| *** belmoreira has joined #openstack-meeting | 11:47 | |
| *** belmoreira has joined #openstack-meeting | 11:48 | |
| *** rh-jelabarre has joined #openstack-meeting | 11:54 | |
| *** rcernin has quit IRC | 12:11 | |
| *** ttsiouts has quit IRC | 12:12 | |
| *** rpittau|bbl is now known as rpittau | 12:19 | |
| *** e0ne has quit IRC | 12:23 | |
| *** e0ne has joined #openstack-meeting | 12:23 | |
| *** raildo has joined #openstack-meeting | 12:23 | |
| *** psachin has quit IRC | 12:26 | |
| *** ysandeep|afk is now known as ysandeep | 12:35 | |
| *** ttsiouts has joined #openstack-meeting | 12:42 | |
| *** ttsiouts has quit IRC | 12:48 | |
| *** ociuhandu has quit IRC | 12:48 | |
| *** ociuhandu has joined #openstack-meeting | 12:49 | |
| *** ttsiouts has joined #openstack-meeting | 13:23 | |
| *** ttsiouts has quit IRC | 13:32 | |
| *** ttsiouts has joined #openstack-meeting | 13:32 | |
| *** ttsiouts has quit IRC | 13:49 | |
| *** dklyle has joined #openstack-meeting | 13:55 | |
| *** yamamoto has joined #openstack-meeting | 13:55 | |
| *** mlavalle has joined #openstack-meeting | 14:00 | |
| slaweq | #startmeeting neutron_drivers | 14:00 |
|---|---|---|
| openstack | Meeting started Fri Apr 17 14:00:33 2020 UTC and is due to finish in 60 minutes. The chair is slaweq. Information about MeetBot at http://wiki.debian.org/MeetBot. | 14:00 |
| njohnston | o/ | 14:00 |
| openstack | Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. | 14:00 |
| *** openstack changes topic to " (Meeting topic: neutron_drivers)" | 14:00 | |
| openstack | The meeting name has been set to 'neutron_drivers' | 14:00 |
| slaweq | hi | 14:00 |
| mlavalle | o/ | 14:00 |
| haleyb | hi | 14:00 |
| amotoki | hi | 14:01 |
| ralonsoh | hi | 14:01 |
| slaweq | lets wait few more minutes for yamamoto and amotoki | 14:01 |
| slaweq | ok, hi amotoki :) | 14:01 |
| yamamoto | hi | 14:01 |
| slaweq | hi yamamoto | 14:01 |
| slaweq | ok, so we can start | 14:01 |
| slaweq | #topic RFEs | 14:01 |
| *** openstack changes topic to "RFEs (Meeting topic: neutron_drivers)" | 14:01 | |
| slaweq | we have 3 RFEs for today | 14:02 |
| slaweq | https://bugs.launchpad.net/neutron/+bug/1869129 | 14:02 |
| openstack | Launchpad bug 1869129 in neutron "neutron accepts CIDR in security groups that are invalid in ovn" [Undecided,In progress] - Assigned to Jake Yip (waipengyip) | 14:02 |
| slaweq | we discussed about it last time but we didn't make any agreement on it | 14:02 |
| slaweq | there is my summary of our last discussion in comment #10 and some new comments from owner in comment #11 | 14:03 |
| njohnston | Yeah, I would like to drop my proposal to add an api extension around /0 netmasks - if someone asks for it that's fine but I'd rather not solve a problem that doesn't really exist. | 14:05 |
| slaweq | I was thinking about it too and I'm not sure if we should change behaviour of API with config option (extension) | 14:06 |
| *** ttsiouts has joined #openstack-meeting | 14:06 | |
| amotoki | I missed the last meeitng. IMHO it looks better to continue to accept non-network address with netmask as CIDR not to break existing tools and basically it should be handled in the driver layer. | 14:06 |
| *** ociuhandu has quit IRC | 14:07 | |
| mlavalle | njohnston: +1 | 14:08 |
| njohnston | yeah, I think the idea that the OVN driver code would silently translate the addresses OVN can't handle into addresses it can makes the most sense from a compatibility perspective | 14:08 |
| slaweq | amotoki: I also tend to choose this option and fix it on drivers side, without changes of existing API | 14:08 |
| slaweq | and maybe some doc update to explain it exactly what are the implications of using wrong cidrs | 14:09 |
| *** moguimar has joined #openstack-meeting | 14:09 | |
| amotoki | another idea/compromise which hits my mind is that we can convert such CIDR into a correct network address when stored in DB. it might affect less to existing tools perhaps. | 14:09 |
| slaweq | amotoki: it's what haleyb proposed last time | 14:09 |
| ralonsoh | I don't agree with "hiding" a user input possible mistake | 14:10 |
| ralonsoh | if we consider the current input is correct, then we should store it as-is in the DB | 14:10 |
| *** ociuhandu has joined #openstack-meeting | 14:11 | |
| haleyb | i am flexible, we know we have to at least fix the driver, was thinking if not changing the current code would lead to duplicate SG rules? | 14:11 |
| amotoki | regarding a check for duplicate SG rules, I think we should improve the check. | 14:12 |
| amotoki | Duplicated rules are not what we expect | 14:12 |
| haleyb | ralonsoh: i don't think the input is incorrect, it's in a grey area, it's just that it won't be used as given, that was why i'd prefer changing it | 14:12 |
| njohnston | amotoki: My main concern is that if a user creates a security group rule for 1.2.3.4/24 when they later list the sg rules it should still say 1.2.3.4/24, which means that we have to store it in the DB the way the user submitted it | 14:13 |
| ralonsoh | exactly | 14:13 |
| njohnston | otherwise a user can't be sure the rule they created is actually in the DB | 14:13 |
| slaweq | njohnston: I agree with that | 14:13 |
| haleyb | njohnston: right, that is the only issue, guess noone is using IPv6 enough to notice my change :) | 14:13 |
| amotoki | njohnston: ralonsoh: I understand the concern. | 14:13 |
| ralonsoh | if 1.2.3.4/24 and 1.2.3.10/24 is a duplicate rule in the backend, we need to handle this in there, in the driver | 14:14 |
| slaweq | maybe we could add new attribute to rule, like "cidr" and store "fixed" cidr there | 14:14 |
| amotoki | if the neutron API returns different reponse, it would be confuing. | 14:14 |
| slaweq | so for e.g. 1.2.3.4/24 in this new field we would store 1.2.3.0/24 | 14:14 |
| ralonsoh | or we keep the current API or we modify it, making a rule check | 14:14 |
| ralonsoh | but nothing in the middle | 14:14 |
| amotoki | /24 means we only care the first 24 bits | 14:14 |
| slaweq | this new field would be read-only | 14:14 |
| *** TrevorV has joined #openstack-meeting | 14:15 | |
| njohnston | I think the idea of a 'normalized cidr' is a good one | 14:16 |
| amotoki | +1 | 14:16 |
| ralonsoh | but you rejected this idea last week! | 14:16 |
| mlavalle | that's ok, people can change opinions | 14:16 |
| mlavalle | I do all the time | 14:17 |
| ralonsoh | then cgoncalves can reply | 14:17 |
| * njohnston is very good at changing opinions except when he isn't | 14:17 | |
| slaweq | ralonsoh: sorry, but I probably missed it somewhere last week | 14:17 |
| njohnston | I like the idea of a net enw column because then we have a uniform behavior for all drivers | 14:17 |
| njohnston | so compatibility is maintained | 14:18 |
| slaweq | ralonsoh: but I really don't remember about proposal of adding new field to SG rule | 14:18 |
| haleyb | so then the drivers would get the rule with this normalized cidr, correct? | 14:18 |
| cgoncalves | my view stands from last time this was discussed. I am just a single voice and a Neutron user, not developer | 14:18 |
| ralonsoh | we agreed to keep the API as is and change the problem we have in OVN | 14:18 |
| ralonsoh | that's all | 14:18 |
| cgoncalves | ralonsoh, +1 | 14:18 |
| mlavalle | even | 14:18 |
| ralonsoh | adding a "rule checker" will change the API | 14:18 |
| mlavalle | ptls are allowed to change opinions | 14:18 |
| slaweq | ralonsoh: I agree to not change exisitng API, but adding new field shouldn't be big problem | 14:18 |
| njohnston | yes, this would not change the API except to add an extra field returned | 14:18 |
| ralonsoh | or "hiding" the input rule, normalizing it | 14:18 |
| amotoki | the new field proposed is almost duplicated to the existing cidr field, but it clarifies what neutron actually recognizes. | 14:19 |
| mlavalle | makes the actual behavior visible | 14:19 |
| slaweq | amotoki: yes, it's something like to be "more verbose" to the user - You set 1.2.3.4/24 and we know about it but we will use 1.2.3.0/24 effectively | 14:19 |
| ralonsoh | we can do this on the fly, we don't need to store anything | 14:20 |
| njohnston | and that's the thing - if we are going to normalize things it would be good to not make a secret about it. that will also make very obvious to anyone who is looking that a rule resolves to 0.0.0.0/0 or whatever | 14:20 |
| haleyb | right, when the driver asks we can convert | 14:20 |
| slaweq | yes, we can, and we are doing it indirectly e.g. in iptables driver now (iptables is doing it for us IIUC) | 14:21 |
| slaweq | but with this new field we could be more verbose for users | 14:21 |
| mlavalle | it's not necessarily predictable behavior, but it becomes visible | 14:22 |
| slaweq | so here are possible options for this bug: | 14:25 |
| cgoncalves | to keep in mind: unless the OVN bug is addressed in its driver (i.e. without relying on the normalization from neutron), the OVN bug will continue to exist in stable branches given that the normalization patch wouldn't be backportable | 14:25 |
| slaweq | 1. fix it only on ovn driver, and silently convert cidr from rule to be good for ovn, | 14:25 |
| slaweq | 2. add new api field and use it e.g. in ovn driver - and in fact in other drivers like iptables or ovs we can use this new field too probably | 14:26 |
| slaweq | I think that other possibilities are already rejected, right? | 14:27 |
| cgoncalves | slaweq, is the OVN bug present in stable branches? | 14:27 |
| slaweq | cgoncalves: I'm not sure, but probably yes | 14:27 |
| njohnston | slaweq: yes I think that is what we have narrowed it down to | 14:27 |
| cgoncalves | slaweq, ok. if it is option 2 won't help for stable branches | 14:27 |
| cgoncalves | may I politely ask why don't someone fix the OVN driver and be done with it? | 14:28 |
| yamamoto | was "3. no change in api/db, fix ovn" rejected? | 14:28 |
| slaweq | cgoncalves: for stable branches we can fix it with silent conversion on backend side | 14:28 |
| *** e0ne has quit IRC | 14:28 | |
| cgoncalves | yamamoto, +1000000 | 14:28 |
| slaweq | yamamoto: it's point 1) from my summary | 14:29 |
| haleyb | i think 1 is two parts - 1a) fix the OVN driver (required), and 1b) change the API/RPC to give the normalized CIDR to drivers | 14:29 |
| yamamoto | "silently convert" in ovn? | 14:29 |
| haleyb | 1a can be backported easily | 14:30 |
| slaweq | ok, so I think I wasn't clear, by silently convert I meant convertion on ovn driver's side to apply what it can appect | 14:30 |
| slaweq | sorry for not being clear | 14:30 |
| slaweq | my 1) is exactly what yamamoto proposed | 14:30 |
| yamamoto | ok | 14:30 |
| haleyb | slaweq: +1 on that, which is what the iptables driver does for a few things | 14:30 |
| *** ttsiouts has quit IRC | 14:31 | |
| *** ysandeep is now known as ysandeep|away | 14:31 | |
| cgoncalves | slaweq, ok, I agree with you. it is the same yamamoto suggested | 14:31 |
| njohnston | I don't know if the openvswitch developers will be willing to fix OVN to accept what they may consider to be "incorrect" CIDRs, but if they do then we also don't know how long until that code is released and what customers may be using it. | 14:31 |
| njohnston | So that is why we are working around the OVN behavior in the neutron OVN driver by doing a conversion. The only question is: do we be transparent tot he user about what is happening? | 14:31 |
| slaweq | njohnston: yes, so IMO we can do it 2 steps: | 14:32 |
| mlavalle | transparency is always the best policy IMO | 14:32 |
| slaweq | 1. fix in ovn driver only - and backport this fix to stable branches | 14:32 |
| slaweq | 2. add api extension to add this normalized cidr field to SG rule | 14:33 |
| njohnston | +100 | 14:33 |
| slaweq | that will not be backported ofcourse | 14:33 |
| mlavalle | in fact transparency should always be baked in our policies | 14:33 |
| amotoki | totally agree | 14:33 |
| *** e0ne has joined #openstack-meeting | 14:34 | |
| amotoki | 2nd step would also give us the clear understanding on how drivers should behave. | 14:34 |
| amotoki | * to driver developers | 14:34 |
| slaweq | so lets vote if that 2 steps solution would be good to go with :) | 14:35 |
| njohnston | +1 | 14:35 |
| amotoki | +1 | 14:35 |
| mlavalle | I am good | 14:35 |
| haleyb | +1 | 14:36 |
| *** ttsiouts has joined #openstack-meeting | 14:36 | |
| slaweq | yamamoto: any thoughts? | 14:37 |
| yamamoto | ok for me, but i'm afraid the new field can confuse users even more. after all, the problem, if any, belongs to doc, IMO. | 14:38 |
| cgoncalves | yamamoto, +1 | 14:38 |
| slaweq | ok, so I will sumup in bug what we agreed here and I will mark it as approved RFE | 14:39 |
| slaweq | thx for great discussion | 14:39 |
| slaweq | next one | 14:40 |
| slaweq | https://bugs.launchpad.net/neutron/+bug/1870319 | 14:40 |
| openstack | Launchpad bug 1870319 in neutron "[RFE] Network cascade deletion API call" [Wishlist,Triaged] - Assigned to Slawek Kaplonski (slaweq) | 14:40 |
| slaweq | this was in fact proposed by dulek from Kuryr team and we talked about it in Shanghai IIRC | 14:40 |
| amotoki | IIRC the main problem is the number of API calls and most of them are related to port deletion. | 14:42 |
| slaweq | I will probably need to write spec with description of all possible resources which should be deleted together with such force network delete | 14:42 |
| slaweq | amotoki: yes, that's the problem for Kuryr basically | 14:42 |
| amotoki | roughly speaking, there seems two ways: (1) cascade network deletion (2) bulk port delete | 14:43 |
| slaweq | as they need to make many API calls to cleanup | 14:43 |
| amotoki | (2) would reduce most API calls | 14:43 |
| amotoki | of course the cascade network deletion would be more friendly to API consumers, but it might be more complicated than bulk port deletion. | 14:44 |
| slaweq | seems that You are right, I can ask dulek if bulk port deletion would solve their problem | 14:44 |
| njohnston | ++ | 14:46 |
| slaweq | I pinged dulek, maybe he will join us soon | 14:46 |
| amotoki | both options have a problem that what response code should return when we hit partial deletion failure. | 14:47 |
| *** jamesmcarthur has quit IRC | 14:47 | |
| slaweq | amotoki: I think that nova supports bulk vms deletion, no? | 14:47 |
| *** jamesmcarthur has joined #openstack-meeting | 14:47 | |
| amotoki | I am not sure at the moment | 14:47 |
| *** ttsiouts has quit IRC | 14:48 | |
| amotoki | it looks like "nova" CLI supports it but the API does not. | 14:48 |
| slaweq | ahh, ok | 14:48 |
| *** dulek has joined #openstack-meeting | 14:49 | |
| slaweq | hi dulek :) | 14:49 |
| dulek | o/ | 14:49 |
| slaweq | we are talking about https://bugs.launchpad.net/neutron/+bug/1870319 | 14:49 |
| openstack | Launchpad bug 1870319 in neutron "[RFE] Network cascade deletion API call" [Wishlist,Triaged] - Assigned to Slawek Kaplonski (slaweq) | 14:49 |
| dulek | Right! So what's the problem? Need any info about use case we have in Kuryr? | 14:49 |
| slaweq | and we would like to know if bulk port deletion would solve Kuryr biggest issue related to this cleanup | 14:49 |
| dulek | Ha, the answer is most likely not what you'd wanted to hear - not really. | 14:50 |
| *** ttsiouts has joined #openstack-meeting | 14:50 | |
| dulek | Or partially. | 14:50 |
| dulek | This is because when we're cleaning up a network in Kuryr, ports are subports of a trunk. | 14:51 |
| dulek | So we cannot delete them before removing them from trunk. | 14:51 |
| dulek | So bulk delete would improve stuff a bit, but we'd still get multiple calls here. | 14:51 |
| dulek | If I'm not mistaken we need to remove them from a trunk one-by-one? | 14:52 |
| amotoki | IIUC we need to delete subport one-by-one now | 14:52 |
| njohnston | any bulk port deletion solution would potentially have to account for some of the ports being trunk ports anyway | 14:52 |
| amotoki | I see. even if we implement bulk port delete, it should be cascade port deletetion | 14:53 |
| dulek | njohnston: True, but currently delete on a port that's a subport will fail. | 14:53 |
| dulek | So if bulk-delete would work the same, we would still need to do n calls, where n is number of ports. | 14:54 |
| amotoki | In my understanding, the main problem is the number of API calls and most of them are related to port deletion. | 14:56 |
| amotoki | if the bulk port deletion is supported including cascading port deletion (of subports), it would reduce most of the APi calls. | 14:56 |
| dulek | That is true. | 14:57 |
| njohnston | what if we added an option to port deletion that would optionally allow the deletion behavior to include removing the port from the trunk? Then the bulk port deletion could be called with that option supplied for all ports. | 14:57 |
| amotoki | While we haven't investigated which is simpler, the cascade network delete or the bulk port delete with cascade delete, can either of them addresses your problem? | 14:57 |
| dulek | Basically it doesn't matter too much if we're doing 3 calls or 1 call, but if we're doing n + x calls, where n is number of ports. | 14:58 |
| amotoki | of course, I understand the cascade network delete woudl be better for you. | 14:58 |
| dulek | amotoki: Sure, 1 call is better than 3 (bulk port, subnet, network), but it's totally a huge improvement over what we have now. | 14:59 |
| njohnston | if we did the bulk port delete with trunk deletion added, then the only remaining steps would be Detach the subnet from the router and Remove the network | 14:59 |
| dulek | Right, forgot about router operation. Anyway it's still constant number of ops, not linear. | 14:59 |
| amotoki | njohnston: i think so | 15:00 |
| njohnston | yep | 15:00 |
| slaweq | njohnston: true, so probably if we would have bulk port deletion we would be close to cascade network deletion as well | 15:00 |
| slaweq | ok, we are over time now | 15:00 |
| slaweq | we will get back to this one | 15:00 |
| slaweq | thx for attendance | 15:01 |
| mlavalle | o/ | 15:01 |
| amotoki | o/ | 15:01 |
| ralonsoh | bye | 15:01 |
| slaweq | and have a great weekend all :) | 15:01 |
| slaweq | o/ | 15:01 |
| slaweq | #endmeeting | 15:01 |
| *** openstack changes topic to "OpenStack Meetings || https://wiki.openstack.org/wiki/Meetings/" | 15:01 | |
| openstack | Meeting ended Fri Apr 17 15:01:12 2020 UTC. Information about MeetBot at http://wiki.debian.org/MeetBot . (v 0.1.4) | 15:01 |
| openstack | Minutes: http://eavesdrop.openstack.org/meetings/neutron_drivers/2020/neutron_drivers.2020-04-17-14.00.html | 15:01 |
| openstack | Minutes (text): http://eavesdrop.openstack.org/meetings/neutron_drivers/2020/neutron_drivers.2020-04-17-14.00.txt | 15:01 |
| openstack | Log: http://eavesdrop.openstack.org/meetings/neutron_drivers/2020/neutron_drivers.2020-04-17-14.00.log.html | 15:01 |
| *** mlavalle has left #openstack-meeting | 15:01 | |
| *** ociuhandu has quit IRC | 15:10 | |
| *** andrebeltrami has joined #openstack-meeting | 15:12 | |
| *** belmoreira has quit IRC | 15:12 | |
| *** ociuhandu has joined #openstack-meeting | 15:13 | |
| *** ricolin has quit IRC | 15:19 | |
| *** moguimar has quit IRC | 15:24 | |
| *** moguimar has joined #openstack-meeting | 15:26 | |
| *** ociuhandu has quit IRC | 15:40 | |
| *** ociuhandu has joined #openstack-meeting | 15:42 | |
| *** ociuhandu has quit IRC | 15:47 | |
| *** jamesmcarthur has quit IRC | 15:58 | |
| *** jamesmcarthur has joined #openstack-meeting | 15:59 | |
| *** dustinc has joined #openstack-meeting | 16:00 | |
| *** e0ne has quit IRC | 16:04 | |
| *** e0ne_ has joined #openstack-meeting | 16:04 | |
| *** ociuhandu has joined #openstack-meeting | 16:05 | |
| *** yamamoto has quit IRC | 16:06 | |
| *** yamamoto has joined #openstack-meeting | 16:07 | |
| *** yamamoto has quit IRC | 16:08 | |
| *** yamamoto has joined #openstack-meeting | 16:08 | |
| *** yamamoto has quit IRC | 16:08 | |
| *** yamamoto has joined #openstack-meeting | 16:09 | |
| *** yamamoto has quit IRC | 16:09 | |
| *** yamamoto has joined #openstack-meeting | 16:09 | |
| *** ttsiouts has quit IRC | 16:12 | |
| *** yamamoto has quit IRC | 16:13 | |
| *** ociuhandu has quit IRC | 16:15 | |
| *** ociuhandu has joined #openstack-meeting | 16:16 | |
| *** rpittau is now known as rpittau|afk | 16:17 | |
| *** ociuhandu has quit IRC | 16:21 | |
| *** jmasud has joined #openstack-meeting | 16:26 | |
| *** ianychoi_ has joined #openstack-meeting | 16:29 | |
| *** armax has quit IRC | 16:29 | |
| *** armax_ has joined #openstack-meeting | 16:30 | |
| *** armax_ is now known as armax | 16:31 | |
| *** gyee has joined #openstack-meeting | 16:31 | |
| *** ianychoi has quit IRC | 16:31 | |
| *** yaawang_ has joined #openstack-meeting | 16:33 | |
| *** yaawang has quit IRC | 16:34 | |
| *** kevinz has quit IRC | 16:40 | |
| *** ociuhandu has joined #openstack-meeting | 16:40 | |
| *** jmasud has quit IRC | 16:55 | |
| *** jmasud has joined #openstack-meeting | 16:57 | |
| *** ociuhandu has quit IRC | 17:03 | |
| *** ttsiouts has joined #openstack-meeting | 17:04 | |
| *** igordc has joined #openstack-meeting | 17:04 | |
| *** yamamoto has joined #openstack-meeting | 17:08 | |
| *** igordc has quit IRC | 17:09 | |
| *** yamamoto has quit IRC | 17:16 | |
| *** alecuyer has left #openstack-meeting | 17:19 | |
| *** links has quit IRC | 17:26 | |
| *** ociuhandu has joined #openstack-meeting | 17:30 | |
| *** e0ne_ has quit IRC | 17:38 | |
| *** e0ne has joined #openstack-meeting | 17:38 | |
| *** jmasud has quit IRC | 17:43 | |
| *** ociuhandu has quit IRC | 17:43 | |
| *** jmasud has joined #openstack-meeting | 17:48 | |
| *** e0ne_ has joined #openstack-meeting | 17:49 | |
| *** e0ne has quit IRC | 17:49 | |
| *** ralonsoh has quit IRC | 17:53 | |
| *** jmasud has quit IRC | 18:05 | |
| *** jamesmcarthur has quit IRC | 18:07 | |
| *** jamesmcarthur has joined #openstack-meeting | 18:08 | |
| *** ociuhandu has joined #openstack-meeting | 18:09 | |
| *** jamesmcarthur has quit IRC | 18:13 | |
| *** ttsiouts has quit IRC | 18:14 | |
| *** ociuhandu has quit IRC | 18:22 | |
| *** darvon has quit IRC | 18:23 | |
| *** ociuhandu has joined #openstack-meeting | 18:23 | |
| *** darvon has joined #openstack-meeting | 18:24 | |
| *** ociuhandu has quit IRC | 18:28 | |
| *** jmasud has joined #openstack-meeting | 18:30 | |
| *** Lucas_Gray has joined #openstack-meeting | 18:38 | |
| *** jamesmcarthur has joined #openstack-meeting | 18:42 | |
| *** e0ne_ has quit IRC | 18:56 | |
| *** e0ne has joined #openstack-meeting | 18:56 | |
| *** jamesmcarthur has quit IRC | 19:00 | |
| *** jamesmcarthur has joined #openstack-meeting | 19:01 | |
| *** jamesmcarthur has quit IRC | 19:15 | |
| *** moguimar has quit IRC | 19:22 | |
| *** ttsiouts has joined #openstack-meeting | 19:26 | |
| *** e0ne has quit IRC | 19:26 | |
| *** raildo has quit IRC | 19:30 | |
| *** ociuhandu has joined #openstack-meeting | 19:35 | |
| *** ociuhandu has quit IRC | 19:42 | |
| *** ttsiouts has quit IRC | 19:47 | |
| *** ttsiouts has joined #openstack-meeting | 19:47 | |
| *** raildo has joined #openstack-meeting | 19:50 | |
| *** jamesmcarthur has joined #openstack-meeting | 20:02 | |
| *** TrevorV has quit IRC | 20:04 | |
| *** jmasud has quit IRC | 20:27 | |
| *** dustinc has quit IRC | 20:29 | |
| *** ociuhandu has joined #openstack-meeting | 20:45 | |
| *** ociuhandu has quit IRC | 21:06 | |
| *** ociuhandu has joined #openstack-meeting | 21:08 | |
| *** rfolco has quit IRC | 21:11 | |
| *** ociuhandu has quit IRC | 21:13 | |
| *** yamamoto has joined #openstack-meeting | 21:14 | |
| *** jmasud has joined #openstack-meeting | 21:15 | |
| *** yamamoto has quit IRC | 21:18 | |
| *** jmasud has quit IRC | 21:29 | |
| *** jmasud has joined #openstack-meeting | 21:31 | |
| *** rh-jelabarre has quit IRC | 21:31 | |
| *** raildo has quit IRC | 21:35 | |
| *** jmasud has quit IRC | 21:44 | |
| *** jamesmcarthur has quit IRC | 21:47 | |
| *** ykatabam has quit IRC | 21:55 | |
| *** ttsiouts has quit IRC | 21:56 | |
| *** ykatabam has joined #openstack-meeting | 21:56 | |
| *** jmasud has joined #openstack-meeting | 22:02 | |
| *** ykatabam has quit IRC | 22:24 | |
| *** ociuhandu has joined #openstack-meeting | 22:32 | |
| *** Lucas_Gray has quit IRC | 22:32 | |
| *** ociuhandu has quit IRC | 22:38 | |
| *** ttsiouts has joined #openstack-meeting | 22:43 | |
| *** ttsiouts has quit IRC | 22:52 | |
| *** igordc has joined #openstack-meeting | 23:08 | |
| *** gyee has quit IRC | 23:10 | |
| *** igordc has quit IRC | 23:27 | |
| *** ykatabam has joined #openstack-meeting | 23:52 | |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!