Krenair | http://developer.openstack.org/api-ref/dns/?expanded=create-zone-transfer-accept-detail lists a zone_id request field | 00:03 |
---|---|---|
Krenair | but then gives a 'zone_transfer_request_id' in the example | 00:03 |
*** johnbelamaric has joined #openstack-dns | 00:18 | |
*** johnbelamaric has quit IRC | 00:22 | |
andrewbogott | I have things in my designate.conf like | 00:24 |
andrewbogott | https://www.irccloud.com/pastebin/VrP3il9e/ | 00:24 |
andrewbogott | But now that I've switched to liberty, the axfr requests just hit 127.0.0.1 and ignore that config. Has that section changed names, perchance? | 00:24 |
*** brad[] has quit IRC | 00:28 | |
*** mlavalle has quit IRC | 00:54 | |
Krenair | The POST path given in the section about accepting zone ownership transfers is also wrong, there's no {zone_id}/ | 00:58 |
*** ducttape_ has joined #openstack-dns | 00:58 | |
Krenair | I tried to upload a fix for it to gerrit, but your CLA and contact information requirement (My real address? That's way, way over the line.) prevented that. I didn't have such issues last time I tried uploading to your gerrit instance a few years ago | 01:02 |
*** stanzgy has joined #openstack-dns | 01:19 | |
*** ducttape_ has quit IRC | 01:39 | |
*** ducttape_ has joined #openstack-dns | 01:39 | |
*** ducttape_ has quit IRC | 01:44 | |
*** richm has quit IRC | 02:07 | |
*** EricGonczer_ has joined #openstack-dns | 02:10 | |
*** ducttape_ has joined #openstack-dns | 02:33 | |
*** EricGonczer_ has quit IRC | 02:39 | |
*** rudrajit has joined #openstack-dns | 02:55 | |
*** ducttape_ has quit IRC | 02:56 | |
*** rudrajit has quit IRC | 03:01 | |
*** ducttape_ has joined #openstack-dns | 03:06 | |
*** ducttape_ has quit IRC | 03:11 | |
*** brad[] has joined #openstack-dns | 03:30 | |
*** EricGonczer_ has joined #openstack-dns | 03:35 | |
*** EricGonczer_ has quit IRC | 03:37 | |
openstackgerrit | Swapnil Kulkarni (coolsvap) proposed openstack/python-designateclient: Remove discover from test-requirements https://review.openstack.org/345759 | 03:55 |
*** j4964d has quit IRC | 04:00 | |
*** rudrajit has joined #openstack-dns | 04:04 | |
*** rudrajit has quit IRC | 04:04 | |
*** rudrajit has joined #openstack-dns | 04:05 | |
*** ducttape_ has joined #openstack-dns | 04:12 | |
*** ducttape_ has quit IRC | 04:17 | |
*** ircuser-1 has joined #openstack-dns | 04:34 | |
*** chrido has joined #openstack-dns | 05:43 | |
*** chrido has quit IRC | 05:49 | |
*** nkinder has quit IRC | 05:49 | |
*** rudrajit has quit IRC | 05:54 | |
*** rudrajit has joined #openstack-dns | 05:54 | |
*** rudrajit has quit IRC | 05:59 | |
*** sticker has quit IRC | 06:34 | |
openstackgerrit | Swapnil Kulkarni (coolsvap) proposed openstack/designate-dashboard: Remove discover from test-requirements https://review.openstack.org/345866 | 06:53 |
*** pcaruana has joined #openstack-dns | 07:18 | |
*** openstackgerrit has quit IRC | 08:11 | |
*** baffle has quit IRC | 08:12 | |
*** Anonissimus has quit IRC | 08:12 | |
*** Anonissimus has joined #openstack-dns | 08:12 | |
*** baffle has joined #openstack-dns | 08:13 | |
*** openstackgerrit has joined #openstack-dns | 08:14 | |
*** ChanServ sets mode: +v openstackgerrit | 08:14 | |
*** nyechiel has joined #openstack-dns | 09:04 | |
*** ducttape_ has joined #openstack-dns | 09:17 | |
*** nyechiel has quit IRC | 09:20 | |
*** ducttape_ has quit IRC | 09:22 | |
*** nyechiel has joined #openstack-dns | 09:38 | |
*** ducttape_ has joined #openstack-dns | 10:18 | |
*** ducttape_ has quit IRC | 10:23 | |
*** nyechiel has quit IRC | 10:36 | |
*** nyechiel has joined #openstack-dns | 11:03 | |
*** stanzgy has quit IRC | 11:04 | |
*** ducttape_ has joined #openstack-dns | 11:08 | |
*** sonuk has quit IRC | 11:14 | |
*** sonuk has joined #openstack-dns | 11:15 | |
*** ducttape_ has quit IRC | 11:19 | |
*** nyechiel has quit IRC | 11:19 | |
*** ducttape_ has joined #openstack-dns | 11:53 | |
*** ducttape_ has quit IRC | 11:57 | |
*** nyechiel has joined #openstack-dns | 12:02 | |
*** nyechiel has quit IRC | 12:56 | |
*** richm has joined #openstack-dns | 12:57 | |
*** krotscheck is now known as lolscheck | 13:12 | |
*** lolscheck is now known as krotscheck | 13:31 | |
*** EricGonczer_ has joined #openstack-dns | 13:42 | |
*** EricGonczer_ has quit IRC | 14:06 | |
andrewbogott | I'm upgrading from kilo to liberty and pdns is no longer getting axfr updates. Logfile looks like this: | 14:21 |
andrewbogott | https://www.irccloud.com/pastebin/gDK3CT1Q/ | 14:21 |
andrewbogott | Can anyone suggest what that 'got lower serial for…' error means? | 14:22 |
*** mlavalle has joined #openstack-dns | 14:27 | |
*** Kiall has joined #openstack-dns | 14:30 | |
timsim | That means it tried to send notifies to 127.0.0.1:53 and nothing was running there (probably), then since it couldn't notify, the zone is out of date. | 14:33 |
*** hwcomcn has joined #openstack-dns | 14:33 | |
*** haplo37_ has joined #openstack-dns | 14:33 | |
andrewbogott | timsim: why is it trying to contact 127.0.0.1? My pool_nameserver is at 208.80.153.48 | 14:34 |
andrewbogott | It doesn't shock me that it's unable to update 127.0.0.1 since there's nothing listening there | 14:34 |
*** hwcomcn has quit IRC | 14:35 | |
timsim | It tries to notify the pool_target, and poll the pool_nameserver | 14:35 |
*** hwcomcn has joined #openstack-dns | 14:35 | |
andrewbogott | timsim: hang on, that's very different from my understanding | 14:36 |
andrewbogott | I thought that pool_target was the pdns database, where new domains were written | 14:36 |
andrewbogott | whereas the pool_nameserver is a running pdns which receives axfr notifications | 14:37 |
andrewbogott | (In any case, both are clearly marked as 208.80.153.48 in my config) | 14:37 |
*** hwcomcn has quit IRC | 14:38 | |
timsim | I think the pool_target options for a database connection should be the pdns database, but the host/port should be wherever you want NOTIFYs to go | 14:39 |
andrewbogott | yep, ok | 14:41 |
timsim | You might need to have [pool_target:xxx] options= host: ip, port: # | 14:41 |
andrewbogott | anyway, here's my config: https://phabricator.wikimedia.org/P3552 | 14:41 |
andrewbogott | That's all in there, as you will see | 14:41 |
andrewbogott | unless the args were renamed between kilo and liberty | 14:41 |
timsim | Yeah I think you need host/port in the "options" line of the pool_target bit | 14:41 |
*** rudrajit has joined #openstack-dns | 14:42 | |
andrewbogott | you mean like options = connection: mysql://pdns:<redacted>@208.80.153.48/pdns, host: 208.80.153.48, port: 53 | 14:44 |
timsim | correct | 14:45 |
timsim | I know it's awful | 14:45 |
andrewbogott | …why... | 14:45 |
timsim | I'm sorry | 14:45 |
andrewbogott | ... | 14:45 |
andrewbogott | No doubt that's written up in the release notes? | 14:45 |
timsim | ¯\_(ツ)_/¯ | 14:45 |
timsim | If that changed between releases, I'm sure it wasn't intentional | 14:46 |
timsim | (That may not even be the fix) | 14:46 |
timsim | (I'm just guessing) | 14:46 |
andrewbogott | so to pick up that change I'd restart the pool manager and mdns, right? | 14:48 |
andrewbogott | Or is central involved here? | 14:49 |
timsim | Just pool manager. | 14:49 |
*** EricGonczer_ has joined #openstack-dns | 14:49 | |
*** nkinder has joined #openstack-dns | 14:51 | |
andrewbogott | timsim: that seems to have done it | 14:56 |
andrewbogott | Does that mean that the other host/port settings in that block are no longer needed? | 14:56 |
timsim | Yep | 14:59 |
andrewbogott | And I assume that block in the config changes back to the other way in M | 14:59 |
timsim | Actually, it changes to a YAML file | 15:00 |
andrewbogott | Every time I upgrade my cluster I spend 10 minutes on each of glance, keystone and nova and then multiple days on designate. | 15:00 |
andrewbogott | That's not just me, is it? | 15:00 |
timsim | What am I supposed to say here? | 15:01 |
andrewbogott | Sorry, just need to blow off steam :) | 15:01 |
timsim | Well, you're welcome for the free support, anyway. | 15:01 |
timsim | It sucks that we somehow broke it that way | 15:02 |
andrewbogott | so — I'm sorry, I don't mean to be ranting at /you/, since you're clearly being helpful and nice :) | 15:02 |
timsim | It's fine. There's no way someone should be expected to just figure that out. | 15:03 |
timsim | We fixed it like 9 months ago, so it sucks to still have to deal with it now | 15:03 |
timsim | Give it another couple cycles and we'll be gone with that nasty config for good | 15:04 |
*** EricGonczer_ has quit IRC | 15:08 | |
timsim | re: Nova/Glance/whatever. Those projects are much older, have many more users, less things change out of utility, and necessity. We're trying to make our project as good as possible, and unlike the Novas/Glances of the world, we have a bit more freedom in fixing our mistakes, rather than being forced to live with so that we don't break people. I'd rather have our relatively small user pool have upgrade pains a few times, but | 15:08 |
timsim | eventually get to a really nice, stable software state. Than have everyone's stuff work perfect between upgrades all the time, but the actual software be shit. | 15:08 |
andrewbogott | Yeah, I understand the philosphy. There's natural tension between 'support existing users' and 'make things better for future users' | 15:10 |
timsim | If we _really_ had the freedom to fix things, we could probably do it all in once cycle, have one painful upgrade, and then we could all go home and just maintain the stable project. But because of the glacial pace imposed by all the red tape you've got to deal with to develop in OpenStack, and the limited resources by nature of all of us having company loyalties, the user often gets a crappy product. Which sucks. | 15:10 |
timsim | It's not the spirit of open source, if we all got to develop on this project out of love, rather than money, we could ship something much better, and ironically, we'd probably care even more about the user experience than the amount we're forced to care by openstack governance. | 15:11 |
timsim | But alas, this is how it works. | 15:11 |
timsim | Sorry, just need to blow off steam :) | 15:11 |
Krenair | Sounds like a breaking change which someone forgot to document | 15:14 |
andrewbogott | Yeah, the change itself isn't a big deal | 15:14 |
andrewbogott | But, I guess more changes = higher chance of them getting left out of the release notes | 15:14 |
mugsie | yeah. | 15:15 |
andrewbogott | Anyway, this works for now, I'll be back to rant again when I upgrade to M :) | 15:15 |
mugsie | we are trying to avoid it as much as possible, and we have added a new grenade gate, that will ensure that upgrades are better | 15:15 |
mugsie | it will test the patch against the default config from the last stable release | 15:15 |
mugsie | which should help | 15:16 |
andrewbogott | ah, great — that would definitely have caught this one | 15:16 |
mugsie | yeah | 15:16 |
mugsie | i will say L->M is going to be a bit of a pain | 15:16 |
mugsie | but there is docs for it | 15:16 |
timsim | Hah, my fault: https://github.com/openstack/designate/commit/520c32e514f0ae611a9875721d0ceeef2b895f38#diff-cb900d2d76e19f687ee632e0edfab0aeR58 | 15:17 |
timsim | !dm timsim | 15:17 |
openstack | timsim: Error: "dm" is not a valid command. | 15:17 |
openstackgerrit | Graham Hayes proposed openstack/designate: Documentation fix up https://review.openstack.org/346077 | 15:23 |
mugsie | timsim: elarson ^^ | 15:25 |
mugsie | Krenair: push patch for your api docs bug | 15:26 |
mugsie | sorry about the CLA, but our hands are tied | 15:26 |
mugsie | the foundationm requires it :( | 15:27 |
mugsie | we used to have some projects that would not enforce the CLA, but all projects have to now | 15:27 |
Krenair | ICLA contributor agreement requires current contact information. | 15:28 |
Krenair | Please review your contact information: | 15:28 |
Krenair | https://review.openstack.org/#/settings/contact | 15:28 |
Krenair | when I try to upload anything | 15:29 |
Krenair | But then you go to the link, put some text in the contact fields and you get "Cannot store contact information" | 15:30 |
mugsie | fwiw the question about "is it needed" comes up every now and again, but we seem to be stuckl with it | 15:31 |
mugsie | http://docs.openstack.org/infra/manual/developers.html#account-setup | 15:31 |
mugsie | you need to go to https://review.openstack.org/#/settings/agreements and click "new contribuitor agreement" | 15:32 |
Krenair | Did that already | 15:32 |
Krenair | mugsie, ... you have to join the foundation? | 15:32 |
mugsie | yeah | 15:32 |
* mugsie doesnt make the rules ..... unfortunatly | 15:33 | |
Kiall | fortunatly, for the rest of us ;) | 15:33 |
Krenair | This is not how open source projects are supposed to work | 15:34 |
mugsie | eh - I would be an amazing linus-esque BDFL | 15:34 |
mugsie | Krenair: OpenStack is different to a lot of Open Source alright. | 15:34 |
mugsie | a lot of projects in this space do ask for the same requriemetns though. | 15:35 |
Krenair | So even if I did make gerrit accept contact details somehow, I'd then be told I need to join an organisation in order to contribute? | 15:35 |
Kiall | Krenair: there's nothing inherentally incompatible with a strict ICLA/CCLA system combined with Open Source, but yea, it certainly puts some folks off - and others give up trying to find the right dance needed to get everything ticked off.. | 15:35 |
mugsie | the contact details error is about the agreement, not the actual details AFAIK | 15:36 |
mugsie | gerrit just considers it a part of your contact details | 15:37 |
Kiall | Krenair: the foundation page is actually where the contact details come from, I believe it's a quirk of Gerrit's CLA system combined with the OpenStack CLA system that it shows the edit button in gerrit at all | 15:37 |
elarson | fwiw, emacs requires you to send in formal copyright assignment to the fsf via fax or mail :) | 15:37 |
Krenair | It already lists a verified ICLA under agreements | 15:37 |
Kiall | elarson: yea, LOTS of open source projects have a CLA requirement. | 15:37 |
Kiall | Krenair: anyway! We're not the best folks to be able to help, other than point you at the docs mugsie already did.. The folks in #openstack-infra run Gerrit etc | 15:38 |
mugsie | yeah, even if the legal standing of such a document is ... debated | 15:38 |
Krenair | I won't be contributing in future | 15:39 |
*** rudrajit has quit IRC | 15:39 | |
*** baffle has quit IRC | 15:39 | |
*** brad[] has quit IRC | 15:39 | |
*** greghaynes has quit IRC | 15:39 | |
*** amit213 has quit IRC | 15:39 | |
*** stupidnic has quit IRC | 15:39 | |
*** federico3 has quit IRC | 15:39 | |
*** krotscheck has quit IRC | 15:39 | |
*** f1ller- has quit IRC | 15:39 | |
*** dasm has quit IRC | 15:39 | |
*** mugsie|alt has quit IRC | 15:39 | |
mugsie | Krenair: that is unfortunate, but please, still feel free to send bugs in, even if you can't push code. | 15:40 |
*** krotscheck has joined #openstack-dns | 15:47 | |
*** mugsie|alt has joined #openstack-dns | 15:47 | |
*** stupidnic has joined #openstack-dns | 15:47 | |
*** baffle has joined #openstack-dns | 15:47 | |
*** brad[] has joined #openstack-dns | 15:47 | |
*** dasm has joined #openstack-dns | 15:47 | |
*** rudrajit has joined #openstack-dns | 15:47 | |
*** f1ller has joined #openstack-dns | 15:47 | |
*** federico3 has joined #openstack-dns | 15:47 | |
*** greghaynes has joined #openstack-dns | 15:48 | |
*** amit213 has joined #openstack-dns | 15:48 | |
*** pcaruana has quit IRC | 15:49 | |
openstackgerrit | Merged openstack/python-designateclient: Remove discover from test-requirements https://review.openstack.org/345759 | 16:31 |
openstackgerrit | Merged openstack/designate-dashboard: Remove discover from test-requirements https://review.openstack.org/345866 | 16:43 |
*** rudrajit has quit IRC | 16:56 | |
*** rudrajit has joined #openstack-dns | 16:56 | |
openstackgerrit | Merged openstack/designate: Documentation fix up https://review.openstack.org/346077 | 16:59 |
*** rudrajit has quit IRC | 17:02 | |
*** dxu has joined #openstack-dns | 17:06 | |
Krenair | So how often is http://developer.openstack.org/api-ref/dns updated from the version in git? | 17:16 |
mugsie | yeah | 17:22 |
mugsie | it should be done anything from 30 -40 mins after | 17:22 |
mugsie | a commit merges * | 17:22 |
*** EmilienM has left #openstack-dns | 17:22 | |
mugsie | there is a largish "post jobs" queue currently | 17:24 |
mugsie | so might be a fewe hours | 17:25 |
Krenair | oh right, it's done as updates go through, not on release | 17:33 |
Krenair | cool | 17:33 |
mugsie | yeah | 17:33 |
Krenair | looking at http://zuul.openstack.org/ some items in that queue have been waiting several hours now | 17:35 |
*** rudrajit has joined #openstack-dns | 17:36 | |
mugsie | yeah, the post jobs are the lowest of the low prioroty. | 17:42 |
mugsie | so they usually back up, and then clear over night, when there is less CI jobs taking build nodes | 17:42 |
*** haplo37_ has quit IRC | 18:00 | |
Krenair | when it 'over night' in openstack? | 18:19 |
mugsie | US night usually :) | 18:25 |
mugsie | us EU people get some good work done in the morning because of that | 18:25 |
*** ducttape_ has joined #openstack-dns | 18:37 | |
*** pglass has joined #openstack-dns | 18:41 | |
mugsie | pglass: if you want you CI to have voting rights on designate, me or kiall can add them | 19:08 |
mugsie | your gerrit account looks like the name is in the wrong format | 19:08 |
*** EricGonczer_ has joined #openstack-dns | 19:10 | |
pglass | well it needs to work first | 19:11 |
pglass | i just started looking at third party ci yesterday | 19:11 |
pglass | so... i didn't think the username had to be in any particular format. just the full name. | 19:13 |
mugsie | oh, maybe | 19:21 |
*** EricGonczer_ has quit IRC | 19:33 | |
FransUrbo | What does "Warning: The quotas could not be retrieved" in Horizon mean? As in, how do I set them (and to what)? | 19:48 |
FransUrbo | .. There's no 'quota-list' option. | 19:49 |
mugsie | FransUrbo: what client are you using? | 19:50 |
mugsie | "openstack" or "designate" | 19:51 |
FransUrbo | designate | 19:51 |
FransUrbo | So this is the 'system' quotas? | 19:51 |
mugsie | yeah | 19:51 |
FransUrbo | Looking through the 'default' quota, there's no reference to 'dns'.. | 19:52 |
mugsie | designso designate quota-get <project-id> | 19:52 |
mugsie | FransUrbo: that is a sore point - it will not be shown in the Horizon UI. | 19:52 |
FransUrbo | Not from the 'openstack quota show' either | 19:53 |
mugsie | same issue | 19:53 |
mugsie | openstack dns quota will work in newton | 19:53 |
FransUrbo | Ok. | 19:54 |
mugsie | (it just got finished in the last month) | 19:54 |
FransUrbo | Ah, well. It's just a warning.. | 19:54 |
FransUrbo | Could you have a (very) quick look at https://bugs.launchpad.net/trove/+bug/1605746 ? I list it as an issue with Trove, but considering I'm having 'Designate related' issues in Neutron as well.. | 19:56 |
openstack | Launchpad bug 1605746 in OpenStack DBaaS (Trove) "Failed to create DNS entry for instance b08666ac-1994-4189-839f-8072eae63d00. No IP available" [Undecided,New] | 19:56 |
FransUrbo | https://bugs.launchpad.net/neutron/+bug/1605336 | 19:56 |
openstack | Launchpad bug 1605336 in neutron "Neutron loadbalancer VIP port fails to create" [Undecided,New] | 19:56 |
mugsie | haha | 19:58 |
mugsie | https://review.openstack.org/346221 | 19:58 |
mugsie | (for the second one) | 19:58 |
mugsie | it was an issue in neutron | 19:59 |
FransUrbo | Neat! | 19:59 |
FransUrbo | yeah, I _did_ state that in the initial blob.. | 19:59 |
mugsie | LBaaS seems to skip a few layers when creating ports, and does not call the neutron API | 20:00 |
mugsie | so, it did not have this extra info that seems to be added somewhere along the way | 20:00 |
mugsie | this trove one looks like it could be them as well. That integration was writen years ago | 20:01 |
FransUrbo | I figured as much, but I thought I'd mention Designate in there none the same.. | 20:04 |
mugsie | yeah - from reading the code, it could be the delay in syncing IP addresses from neutron to nove | 20:04 |
mugsie | so neutron gives it the IP, but nova has not synced the change in yet | 20:05 |
mugsie | so when it goes to use the IP, the erorr handling kicks in | 20:05 |
FransUrbo | Sounds resonable. I get that almost immediately. | 20:09 |
mugsie | they are fairly friendly in #openstack-trove - they would know better than me | 20:11 |
FransUrbo | Yeah, but they're never there :D. I mentioned it a couple of hours, waiting for the rock to rotate to the right position :D | 20:11 |
mugsie | ping amrith and tell him I sent you :) | 20:12 |
mugsie | he should be around by now afaik | 20:12 |
FransUrbo | Thanx. | 20:13 |
FransUrbo | I talked to him earlier today and he was quite testy about me not using redstack or whatever to install my Trove installation.. | 20:14 |
FransUrbo | But eventually I got out of him that it (Trove) don't work with Heat.. | 20:14 |
mugsie | as in Heat installing trove, or trove using heat to create instances? | 20:16 |
FransUrbo | the latter | 20:16 |
mugsie | ah, yeah | 20:16 |
mugsie | no | 20:16 |
FransUrbo | I thought it would be nice, to use a specialized service to do the deploy. Not inventing the wheel twice and all that. But apparently it's being removed. | 20:17 |
mugsie | It was added to satisfy a requirement from the TC years ago | 20:17 |
FransUrbo | TC? | 20:18 |
mugsie | it was never the real one used in production | 20:18 |
mugsie | technical committee | 20:18 |
FransUrbo | Ah. | 20:18 |
mugsie | we used to have a process where you would apply to them to be come part of OpenStack | 20:18 |
mugsie | they wouild then review the prohect | 20:18 |
FransUrbo | Ah, well. After disabling that in Trove, everything started to fall into place. Except that DNS issue.. | 20:18 |
mugsie | and they could ask you to do things to fit in better | 20:18 |
FransUrbo | Yeah, I saw your discussion earlier.. | 20:19 |
mugsie | OK. they should be looking at moving it to the neutron one anyway, so I will ping them again during the week | 20:19 |
FransUrbo | Neutron?! | 20:19 |
mugsie | so they should just be sending the name to neutron when they create the port | 20:20 |
FransUrbo | Got it.. | 20:20 |
mugsie | then neutron can set it in designate, and ensure it is deleted when the port is deleted | 20:20 |
FransUrbo | ... and Neutron then talks to Designate? | 20:20 |
mugsie | yeah - that was the bit that was failing for the LBaaS bug | 20:21 |
FransUrbo | Yeah. I'm building my own local package right now to test it. | 20:21 |
mugsie | neutron adds records to its inbuild dnsmasq for internal addresses, and for external addresses it pushes them to us | 20:21 |
FransUrbo | Yeah, you told me that yesterday (or the day before) :D | 20:24 |
mugsie | ah - I am foirgeting things in my old age :) | 20:24 |
FransUrbo | ROTFL | 20:24 |
FransUrbo | Know the feeling :) | 20:24 |
FransUrbo | Btw, the VIP fix works. Want me to sign of on it? | 20:32 |
mugsie | please :) | 20:33 |
FransUrbo | Where/how do I do that? | 20:33 |
*** amrith has joined #openstack-dns | 20:34 | |
amrith | mugsie, yo yo | 20:34 |
mugsie | hey FransUrbo ^ is the reporter for that bug :) | 20:34 |
amrith | yes | 20:34 |
amrith | I don't have scrollback on this channel, let me go find eavesdrop | 20:34 |
mugsie | FransUrbo: https://review.openstack.org/346221 | 20:34 |
amrith | hmm, reading scrollback | 20:35 |
FransUrbo | Just saw the Reply.. | 20:36 |
mugsie | amrith: looks like the issue is https://github.com/openstack/trove/blob/master/trove/taskmanager/models.py#L980 | 20:36 |
amrith | first of all, he never come to trove and said mugsie sent him | 20:36 |
mugsie | in that region | 20:36 |
mugsie | :) | 20:36 |
amrith | if he had, we'd have been all over it; like use-car salesmen on a prospective customer | 20:36 |
mugsie | there seems to be an issue with getting the IP from Nova? | 20:36 |
mugsie | sorry - https://github.com/openstack/trove/blob/master/trove/taskmanager/models.py#L1001 | 20:37 |
mugsie | so it is polling, then thinks it has the IP. it trys to access it then, but the IP is not set | 20:37 |
* amrith reads the scrollback with amusement ... http://eavesdrop.openstack.org/irclogs/%23openstack-dns/%23openstack-dns.2016-07-22.log.html#t2016-07-22T20:14:25 | 20:37 | |
FransUrbo | Ok, review added on 346221 | 20:38 |
mugsie | FransUrbo: tnx :) | 20:38 |
FransUrbo | np | 20:38 |
FransUrbo | amrith: Well, you WHERE testy :) | 20:39 |
mugsie | amrith: you should really move to setting the dns_name attribute on ports in neutron - then the data can be pushed through neutron to us | 20:39 |
mugsie | FransUrbo: amrith has been called that from time to time :) | 20:39 |
FransUrbo | LOL | 20:39 |
mugsie | never by me of course :) | 20:39 |
amrith | Were or Where? | 20:39 |
FransUrbo | I didn't take offence. I'm been around to long :) | 20:39 |
amrith | mugsie, if people find me testy, I wonder what they'll think of you. | 20:40 |
FransUrbo | were, dang :) | 20:40 |
FransUrbo | Mugsie soften up after a while.. :) | 20:40 |
* mugsie is making a few waves this week. | 20:41 | |
amrith | so mugsie, I've got a plugin for you :) it is 220v, 50hz | 20:41 |
amrith | ok, I read all the scrollback | 20:42 |
amrith | so what appears to be the problem with DNS? | 20:42 |
mugsie | when the integration is turned on, trove trys to get the IP, can't and goes to ERROR | 20:42 |
FransUrbo | Did you see the issue I reported? | 20:42 |
amrith | I did, and I marked it incomplete | 20:43 |
mugsie | without the integration, the IP is nover looked up(not at that point anyway) | 20:43 |
amrith | I have no idea what you setup or how you got to that | 20:43 |
amrith | integration with what? dns? | 20:43 |
mugsie | designate | 20:43 |
amrith | ok | 20:43 |
amrith | ah yes, https://github.com/openstack/trove/blob/master/trove/taskmanager/models.py#L971 | 20:43 |
mugsie | (trove has designate integration from years back) | 20:43 |
amrith | yup | 20:44 |
amrith | so why aren't we getting an IP? | 20:44 |
mugsie | it was supposed to be deployed in public cloud, but we never got to turn it on :( | 20:44 |
mugsie | self.dns_ip_address is not being set | 20:45 |
amrith | we timeout on getting an IP in ip_is_available after whatever DNS_TIME_OUT happens to be. | 20:45 |
mugsie | timeout hasnt hit | 20:45 |
mugsie | it thinks it has one | 20:45 |
FransUrbo | Included my Trove config options if that helps. | 20:46 |
amrith | timeout is 120s mugsie | 20:46 |
FransUrbo | This is triggered almost right away (well within that 120s). | 20:47 |
mugsie | FransUrbo: can you add in some debuging loggig statements? | 20:47 |
FransUrbo | Sure. | 20:47 |
FransUrbo | Just tell me what to add where.. | 20:48 |
mugsie | just above "LOG.debug("Creating dns entry...")" add "LOG.debug(self.addresses)" | 20:48 |
amrith | mugsie, no need to | 20:48 |
amrith | nova pushed the instance to error state | 20:48 |
amrith | so we raise in ip_is_available() | 20:49 |
amrith | https://github.com/openstack/trove/blob/master/trove/taskmanager/models.py#L993 | 20:49 |
amrith | we won't wait 120s | 20:49 |
amrith | so why did Nova error the instance? | 20:49 |
FransUrbo | Want me to get nova logs? | 20:49 |
mugsie | the error message is from - https://github.com/openstack/trove/blob/master/trove/taskmanager/models.py#L1003 | 20:49 |
mugsie | which is after that popint, and it trggered by self.dns_ip_address being None / False | 20:50 |
mugsie | "No IP available" is only from there | 20:50 |
amrith | well in that case, I see why you want to print self.addresses | 20:51 |
amrith | coz it is returning true at 983 | 20:51 |
mugsie | yup :) | 20:51 |
amrith | hey mugsie, shouldn't 1005 be before 1002? | 20:54 |
amrith | no, never mind | 20:54 |
amrith | I bet the issue is that network_label_regex is wrong, or there's a blacklist | 20:56 |
mugsie | oh. that coud be it | 20:57 |
amrith | see https://github.com/openstack/trove/blob/master/trove/instance/models.py#L214-L227 | 20:57 |
*** f1ller is now known as filler | 20:58 | |
mugsie | yes. FransUrbo do you have network_label_regex set ? | 20:59 |
FransUrbo | yes. network_label_regex = .* | 21:01 |
amrith | mugsie, go get a pint for me mate. | 21:05 |
amrith | I can't have any ... so you have to support the breweries now ... all by your lonesome. | 21:06 |
mugsie | FransUrbo: OK, I am going to leave the office (its 22:00 here) - can you dump your full config into an etherpad / paste? | 21:06 |
mugsie | and also the output from that extra debug log? | 21:06 |
mugsie | I will have a look on monday | 21:06 |
mugsie | amrith: I will do my best :) | 21:07 |
FransUrbo | I'm trying to get you logs, but I'm hitting a quota limit on secgroups. And I can't seem to change it! | 21:07 |
*** kbyrne has quit IRC | 21:07 | |
FransUrbo | Is there any way to tell Trove to use existing SG instead of creating one? | 21:08 |
*** filler is now known as f1ller | 21:08 | |
amrith | mugsie, a thought. it would be interesting to see IPs at https://github.com/openstack/trove/blob/master/trove/instance/models.py#L225 just in case filter is causing an issue. | 21:08 |
amrith | .. | 21:09 |
*** EricGonczer_ has joined #openstack-dns | 21:10 | |
*** f1ller is now known as filler | 21:10 | |
*** kbyrne has joined #openstack-dns | 21:11 | |
FransUrbo | Ok, self.addresses => http://paste.openstack.org/show/539891/ | 21:13 |
FransUrbo | For me, both ip_regex and black_list_regex is not set specifically (and there's no default). | 21:17 |
FransUrbo | So that section would not trigger.. | 21:17 |
FransUrbo | mugsie: My config dump: http://bayour.com/misc/trove-cfgs.txt | 21:26 |
*** ducttape_ has quit IRC | 21:59 | |
*** pglass has quit IRC | 22:12 | |
openstackgerrit | Tyr Johanson proposed openstack/designate-dashboard: [WIP] API v2 Dashboard https://review.openstack.org/341182 | 22:40 |
*** dxu has quit IRC | 22:45 | |
FransUrbo | amrith: Still here? | 23:42 |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!