| *** stanzgy has joined #openstack-dns | 01:34 | |
| *** stanzgy_ has joined #openstack-dns | 01:34 | |
| *** richm has quit IRC | 02:19 | |
| *** mwagner_lap has joined #openstack-dns | 02:35 | |
| *** GonZo2K has quit IRC | 03:28 | |
| *** nkinder has joined #openstack-dns | 03:38 | |
| *** csoukup has joined #openstack-dns | 03:55 | |
| *** andrewbogott_afk is now known as andrewbogott | 04:35 | |
| *** andrewbogott is now known as andrewbogott_afk | 04:41 | |
| *** csoukup has quit IRC | 04:47 | |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/designate: Imported Translations from Transifex https://review.openstack.org/160084 | 06:35 |
|---|---|---|
| *** chlong has quit IRC | 08:02 | |
| zigo_ | andrewbogott_afk: I think it's just best if you rebuild designate from Debian's sources. | 09:01 |
| *** jordanP has joined #openstack-dns | 10:05 | |
| *** stanzgy has quit IRC | 10:53 | |
| *** stanzgy_ has quit IRC | 10:54 | |
| *** untriaged-bot has joined #openstack-dns | 11:02 | |
| untriaged-bot | Untriaged bugs so far: | 11:02 |
| untriaged-bot | https://bugs.launchpad.net/designate/+bug/1416337 | 11:02 |
| openstack | Launchpad bug 1416337 in Designate "Designate server create with concurrent request is not listing all servers even after successful creation." [Undecided,Incomplete] | 11:02 |
| untriaged-bot | https://bugs.launchpad.net/designate/+bug/1425668 | 11:02 |
| openstack | Launchpad bug 1425668 in Designate "Poor error message when using same database for designate and the pool manager cache" [Undecided,New] | 11:02 |
| untriaged-bot | https://bugs.launchpad.net/designate/+bug/1425117 | 11:02 |
| openstack | Launchpad bug 1425117 in Designate "Designate does not work with postgres" [High,New] | 11:02 |
| *** untriaged-bot has quit IRC | 11:02 | |
| openstackgerrit | Kiall Mac Innes proposed openstack/designate: Implement Scoped TSIG Key Support https://review.openstack.org/160165 | 11:59 |
| openstackgerrit | Kiall Mac Innes proposed openstack/designate: WIP: Service Class Cleanup - Part 3/3 https://review.openstack.org/160129 | 12:06 |
| *** GonZo2K has joined #openstack-dns | 12:58 | |
| *** GonZo2K has quit IRC | 12:58 | |
| *** GonZo2K has joined #openstack-dns | 12:58 | |
| *** andrewbogott_afk is now known as andrewbogott | 13:44 | |
| *** fouxm has quit IRC | 13:54 | |
| *** fouxm has joined #openstack-dns | 13:57 | |
| *** richm has joined #openstack-dns | 14:06 | |
| *** nkinder has quit IRC | 14:19 | |
| *** vinod1 has joined #openstack-dns | 14:40 | |
| *** csoukup has joined #openstack-dns | 15:03 | |
| *** vinod1 has quit IRC | 15:13 | |
| *** nkinder has joined #openstack-dns | 15:13 | |
| *** paul_glass has joined #openstack-dns | 15:17 | |
| openstackgerrit | Merged openstack/designate: Imported Translations from Transifex https://review.openstack.org/160084 | 15:37 |
| *** paul_glass has quit IRC | 15:47 | |
| *** jmcbride has joined #openstack-dns | 15:47 | |
| *** vinod1 has joined #openstack-dns | 16:02 | |
| openstackgerrit | Satyanarayana Patibandla proposed openstack/python-designateclient: Added previllege to admin tenant to view other tenants created records. https://review.openstack.org/160397 | 16:21 |
| *** boris-42 has quit IRC | 16:32 | |
| openstackgerrit | Tim Simmons proposed openstack/designate: Agent configuration and BIND9 improvements https://review.openstack.org/158462 | 16:42 |
| *** rmoe has quit IRC | 16:47 | |
| *** paul_glass has joined #openstack-dns | 16:50 | |
| *** jordanP has quit IRC | 17:02 | |
| *** untriaged-bot has joined #openstack-dns | 17:02 | |
| untriaged-bot | Untriaged bugs so far: | 17:02 |
| untriaged-bot | https://bugs.launchpad.net/designate/+bug/1416337 | 17:02 |
| openstack | Launchpad bug 1416337 in Designate "Designate server create with concurrent request is not listing all servers even after successful creation." [Undecided,Incomplete] | 17:02 |
| untriaged-bot | https://bugs.launchpad.net/designate/+bug/1425668 | 17:02 |
| openstack | Launchpad bug 1425668 in Designate "Poor error message when using same database for designate and the pool manager cache" [Undecided,New] | 17:02 |
| untriaged-bot | https://bugs.launchpad.net/designate/+bug/1425117 | 17:02 |
| openstack | Launchpad bug 1425117 in Designate "Designate does not work with postgres" [High,New] | 17:02 |
| *** untriaged-bot has quit IRC | 17:02 | |
| *** jordanP has joined #openstack-dns | 17:03 | |
| *** mugsie has quit IRC | 17:05 | |
| *** mugsie has joined #openstack-dns | 17:06 | |
| *** mugsie is now known as Guest46000 | 17:06 | |
| *** GonZo2K has quit IRC | 17:07 | |
| *** rmoe has joined #openstack-dns | 17:08 | |
| *** Guest46000 has quit IRC | 17:19 | |
| *** Guest46000 has joined #openstack-dns | 17:19 | |
| *** Guest46000 is now known as mugsie_ | 17:19 | |
| Kiall | timsim: about? re https://review.openstack.org/#/c/160127/ - I honestly forgot I removed the constructor "config" option from that one, I think that's the answer to the Q I had left here.. https://review.openstack.org/#/c/160127/2/designate/agent/service.py | 17:22 |
| timsim | Yeah i'm here | 17:23 |
| Kiall | by making "config" part of the standard constructor args, we can just read them from there.. Down the line, I'd like to see the config options be defined DRY-ly .. But, for now I'm aiming to get DNSService used fby both mDNS and Agent, so I can do the TSIG parts of mDNS | 17:23 |
| timsim | Yeah my only thought there was if it might make reloadable config harder | 17:24 |
| Kiall | Speaking of which - The agent and ACL - We don't currently restrict who can send commands in, right? | 17:24 |
| timsim | There's an allow-notify list | 17:24 |
| timsim | That checks every request's source IP. if it's not in the list it refuses it. | 17:24 |
| Kiall | Okay.. So TSIG isn't needed there to identify the caller - which makes my life easier for now :) | 17:25 |
| timsim | Yeah not yet, I was really hoping to add it at some point, as people will definitely want it. But not necessary for now. | 17:25 |
| Kiall | Any other thoughts on the 3 parts of that change? I'll update the config stuff now.. | 17:26 |
| timsim | Nah I didn't have anything else. Looked pretty clean. I haven't tested it out though. | 17:27 |
| Kiall | Cool :) | 17:31 |
| *** boris-42 has joined #openstack-dns | 18:05 | |
| andrewbogott | zigo: did you say that you have Trusty/Icehouse packages for designate someplace? Or just Jessie/Icehouse? (Even the latter would move me forward quite a bit.) | 18:06 |
| openstackgerrit | Graham Hayes proposed openstack/designate: Fixed _is_superdomain check in central https://review.openstack.org/158693 | 18:09 |
| Kiall | andrewbogott: https://packages.debian.org/search?keywords=designate&searchon=names&suite=unstable§ion=all | 18:17 |
| andrewbogott | Kiall: Would you expect those to work on Trusty as well? I landed in dependency hell when I tried to build something similar. | 18:18 |
| Kiall | I suspect dependancy hell can be solved with the icehouse ubuntu cloud archive, and rebuilding the designate package after modifying the debian/control file to match whatever the cloud archive has available.. the versions there should be OK, since we pin to the same versions of deps as the rest of openstack.. | 18:19 |
| Kiall | so even if debian has a newer older, what's in the cloud archive should be good for our icehouse release | 18:20 |
| Kiall | newer or older versions* | 18:20 |
| andrewbogott | There is no icehouse cloud archive for Trusty though… | 18:20 |
| andrewbogott | But, ok, rebuilds will probably get me there. Thanks | 18:20 |
| Kiall | Ohh, for trusty.. that shipped with icehouse, right? If so, then the stock ubuntu archive should have what we need | 18:21 |
| *** mugsie_ is now known as mugsie | 18:22 | |
| Kiall | Anyway! Dependancy wise - we are tied to the versions used by the rest of OpenStack Icehouse, so if nova etc work on it, then so will we (most likely anyway!).. So, updating the debian/control to match what's available should do it.. The only exception would be if the packages use deb packaging features not available on Trusty .. But I'm not really sure if they do/don't etc etc | 18:24 |
| *** jordanP has quit IRC | 18:25 | |
| openstackgerrit | Kiall Mac Innes proposed openstack/designate: Service Class Cleanup - Part 2/3 https://review.openstack.org/160128 | 18:25 |
| openstackgerrit | Kiall Mac Innes proposed openstack/designate: WIP: Service Class Cleanup - Part 3/3 https://review.openstack.org/160129 | 18:25 |
| openstackgerrit | Kiall Mac Innes proposed openstack/designate: Service Class Cleanup - Part 1/3 https://review.openstack.org/160127 | 18:25 |
| Kiall | timsim / mugsie.. updated re the service config discussion we had.. Using a mix of both your suggestions ;) | 18:25 |
| andrewbogott | hm… may be my packaging skills are just not up to the job; I will persevere | 18:25 |
| Kiall | Yea - It's been about a year since I touched debian packaging too ;) | 18:26 |
| openstackgerrit | Kiall Mac Innes proposed openstack/designate: Service Class Cleanup - Part 2/3 https://review.openstack.org/160128 | 18:32 |
| openstackgerrit | Kiall Mac Innes proposed openstack/designate: WIP: Service Class Cleanup - Part 3/3 https://review.openstack.org/160129 | 18:32 |
| openstackgerrit | Kiall Mac Innes proposed openstack/designate: Service Class Cleanup - Part 1/3 https://review.openstack.org/160127 | 18:32 |
| Kiall | Always miss something in the first 3 attempts at a long chain ;) | 18:32 |
| *** jmcbride has quit IRC | 18:42 | |
| *** jmcbride has joined #openstack-dns | 18:52 | |
| *** crc32 has joined #openstack-dns | 18:55 | |
| openstackgerrit | Merged openstack/designate: Add howto for creating PTR records https://review.openstack.org/157995 | 18:59 |
| *** crc32 has quit IRC | 19:03 | |
| *** GonZo2K has joined #openstack-dns | 19:11 | |
| *** GonZo2K has quit IRC | 19:12 | |
| openstackgerrit | Tim Simmons proposed openstack/designate: Agent configuration and BIND9 improvements https://review.openstack.org/158462 | 19:12 |
| *** crc32 has joined #openstack-dns | 19:12 | |
| *** shakamunyi has joined #openstack-dns | 19:20 | |
| Kiall | Great.. pep8 fail in commit 1 of 3.. heh :( | 19:23 |
| * Kiall needs `git rebase` to automatically do a `tox -epy27,pep8` on every commit in the chain -_- | 19:23 | |
| openstackgerrit | Kiall Mac Innes proposed openstack/designate: Service Class Cleanup - Part 2/3 https://review.openstack.org/160128 | 19:24 |
| openstackgerrit | Kiall Mac Innes proposed openstack/designate: Service Class Cleanup - Part 3/3 https://review.openstack.org/160129 | 19:24 |
| openstackgerrit | Kiall Mac Innes proposed openstack/designate: Service Class Cleanup - Part 1/3 https://review.openstack.org/160127 | 19:24 |
| *** GonZo2K has joined #openstack-dns | 19:26 | |
| *** GonZo2K has quit IRC | 19:26 | |
| *** GonZo2K has joined #openstack-dns | 19:26 | |
| elarson | rjrjr: I was wondering if you might be able to help me understand the locking https://github.com/openstack/designate/blob/master/designate/pool_manager/service.py#L235 I'm curious why update has a lock while the other operations don't | 19:39 |
| elarson | rjrjr: I'm also wondering if the lock doesn't really protect against anything if you were running more than one pool manager process | 19:40 |
| Kiall | elarson: I remember a conversation about this at the mid-cycle, but I'm struggeling to remember the exact reasoning.. | 19:44 |
| Kiall | I think the reasoning was that, since N update_statuses can come in at once, that method needs be wrapped in a lock to work at all even with a single pool manager.. And that we'll need something more for reliablity working with multiple active pool managers per pool.. | 19:46 |
| * elarson is reading the code more closely :) | 19:49 | |
| *** crc32 has quit IRC | 19:56 | |
| elarson | Kiall: it looks like, generally, getting a serial consensus across backends could be problematic. that is my best guess at the moment :) | 19:57 |
| elarson | problematic == non-trivial to get a single, canonical answer | 19:58 |
| Kiall | Race condition wise? With a single pool manager, it's doable as our consensus is based on the question "Do you have at least serial #1234" rather than "Do you have exactly serial #1234" | 19:58 |
| Kiall | With multiple pool managers and no lock, things get more interesting with race conditions.. and it may be that we need to lock cross all the pool managers for a particular pool :/ | 19:59 |
| *** johnbelamaric has joined #openstack-dns | 20:07 | |
| *** johnbelamaric1 has joined #openstack-dns | 21:04 | |
| *** johnbelamaric has quit IRC | 21:05 | |
| *** shakamunyi has quit IRC | 21:07 | |
| *** shakamunyi has joined #openstack-dns | 21:08 | |
| timsim | Can we just not upgrade eventlet? | 21:22 |
| *** chlong has joined #openstack-dns | 21:24 | |
| Kiall | timsim: lol, what? | 21:24 |
| timsim | I've got a deployment that's got the latest code, but works with eventlet 0.16.1 and doesn't w/ 0.17.1 | 21:25 |
| Kiall | Ah.. 0.17.0 broke dns.name.Name, and 0.17.1 exposed some issues with our code.. | 21:26 |
| timsim | http://paste.openstack.org/show/185485/ | 21:26 |
| Kiall | (The 2 gate blockers we had in 2 days ;)) | 21:26 |
| Kiall | OO | 21:26 |
| Kiall | Humm - That's different | 21:26 |
| timsim | I mean I knew there were issues. Patched them, thought it was good, but just this notify exchange fails. | 21:27 |
| Kiall | I've not noticed that, is it happening in the gate? or just in your env? | 21:27 |
| timsim | Oh just my crazy env | 21:27 |
| timsim | It's MDNS->Proxy->Agent Notify via TCP | 21:28 |
| Kiall | lol | 21:28 |
| Kiall | Proxy? | 21:28 |
| timsim | Long story. | 21:28 |
| Kiall | UDP Notify certainly working (Just saw one ;)) | 21:28 |
| timsim | It's HAProxy. Yeah they're TCP | 21:28 |
| timsim | And it works fine on old eventlet. | 21:29 |
| Kiall | I suppose they could have broken something else we haven't noticed yet :/ | 21:30 |
| timsim | The odd thing is that my CREATE/DELETE messages go through fine. | 21:30 |
| elarson | s/eventlet/gevent and monkey patch everything? | 21:30 |
| Kiall | I can tell you straight out - s/eventlet/gevent/ will result in whatever DC you're running this in going on fire... | 21:32 |
| elarson | no big deal, just grab a fire extinguisher! problem solved | 21:33 |
| timsim | It seems like the only thing they're patching that we use now is dns.query (which is what's sending the NOTIFY) https://github.com/eventlet/eventlet/commit/3de49aef1399594b2cab0652d9ff3976d19b5e72#diff-47bd8cdaac88cccf21624af4df7a35c3R47 | 21:34 |
| elarson | Kiall: btw, I'd be surprised if we need that locking since we're using an eventlet threadpool | 21:35 |
| elarson | that said, I would *not* be surprised if I were wrong ;) | 21:35 |
| *** andrewbogott is now known as andrewbogott_afk | 21:41 | |
| *** andrewbogott_afk is now known as andrewbogott | 21:43 | |
| Kiall | elarson: at points within the lock, we do I/O, which gives eventlet time to switch to another thread, and quite possibly reenter the same method for another status update.. | 21:44 |
| Kiall | So close -_- | 21:50 |
| Kiall | 2015-03-02 21:50:41.981 DEBUG designate.service [-] Handling UDP Request from: 127.0.0.1:39464 from (pid=32240) _dns_handle_udp /opt/stack/designate/designate/service.py:307 | 21:50 |
| Kiall | 2015-03-02 21:50:41.986 WARNING dns.tsig [-] Keyname: test-key. | 21:50 |
| Kiall | 2015-03-02 21:50:41.987 WARNING dns.tsig [-] Secret: +Cdjlkef9ZTSeixERZ433Q== | 21:50 |
| Kiall | 2015-03-02 21:50:41.988 ERROR designate.dnsutils [-] Invalid TSIG signature from 127.0.0.1:39464 | 21:50 |
| *** chlong has quit IRC | 21:51 | |
| *** shakamunyi_ has joined #openstack-dns | 21:53 | |
| *** shakamunyi has quit IRC | 21:54 | |
| openstackgerrit | Merged openstack/designate: Remove explicit thread-local update in context https://review.openstack.org/160114 | 21:59 |
| *** nkinder has quit IRC | 22:01 | |
| openstackgerrit | Merged openstack/designate: Sync with oslo-incubator https://review.openstack.org/160115 | 22:07 |
| *** johnbelamaric has joined #openstack-dns | 22:08 | |
| openstackgerrit | Merged openstack/designate: Update default managed_resource_email https://review.openstack.org/160171 | 22:09 |
| openstackgerrit | Merged openstack/designate: Add Developer Guidelines docs page https://review.openstack.org/160174 | 22:09 |
| *** johnbelamaric1 has quit IRC | 22:11 | |
| *** GonZoPT has joined #openstack-dns | 22:22 | |
| *** GonZo2K has quit IRC | 22:24 | |
| *** csoukup has quit IRC | 22:40 | |
| Kiall | Warning.. openstackgerrit is about the spam us ;) | 22:43 |
| openstackgerrit | Kiall Mac Innes proposed openstack/designate: Implement Scoped TSIG Key Support https://review.openstack.org/160165 | 22:43 |
| openstackgerrit | Kiall Mac Innes proposed openstack/designate: Service Class Cleanup - Part 2/3 https://review.openstack.org/160128 | 22:43 |
| openstackgerrit | Kiall Mac Innes proposed openstack/designate: Service Class Cleanup - Part 3/3 https://review.openstack.org/160129 | 22:43 |
| openstackgerrit | Kiall Mac Innes proposed openstack/designate: Service Class Cleanup - Part 1/3 https://review.openstack.org/160127 | 22:43 |
| openstackgerrit | Kiall Mac Innes proposed openstack/designate: WIP: Implement TSIG Support in mDNS https://review.openstack.org/160543 | 22:43 |
| paul_glass | vinod1: what behavior should I expect with the noop pool manager cache? | 22:45 |
| *** jmcbride has quit IRC | 22:45 | |
| vinod1 | It is mostly used with unit tests. | 22:46 |
| vinod1 | Theoritically if you use it - it should fetch from the minidns for every access to the cache | 22:47 |
| vinod1 | Though checking for consensus relies on cache entries - so that might not work | 22:48 |
| vinod1 | rjrjr: Thoughts/comments on this? | 22:49 |
| paul_glass | with noop cache, I see zones go to ERROR, which makes sense I guess. | 22:53 |
| Kiall | In theory, they shouldn't go to ERROR.. If they do, it's a bug | 22:56 |
| Kiall | a cache can be wiped at any moment in time, and we should handle that when it happens.. The noop cache let's us simulate loosing the cache all the time.. ;) | 22:57 |
| paul_glass | but if the pool manager doesn't do anything, there's no part of designate that knows if the zone is actually active on the nameserver, is there? | 22:59 |
| paul_glass | how could the zone ver go to active? | 22:59 |
| paul_glass | ever* | 22:59 |
| *** johnbelamaric has quit IRC | 23:03 | |
| paul_glass | oh, it's a noop cache, not a noop pool manager. | 23:04 |
| *** nkinder has joined #openstack-dns | 23:05 | |
| *** shakamunyi_ has quit IRC | 23:06 | |
| Kiall | vinod1 / timsim / mugsie - That chain of 5 patches above is about 98.2% of TSIG support in mDNS etc :) | 23:12 |
| Kiall | Still some final bits to work out, but pretty much there | 23:12 |
| vinod1 | 98.2% - you are really precise :-) | 23:13 |
| Kiall | ;) | 23:13 |
| mugsie | well, its missing the vital bit ;) | 23:13 |
| Kiall | :D | 23:14 |
| rjrjr | vinod1: the noop cache will use mdns, after the zone goes to ERROR. what we talked about at the midcycle meetup was to have the sync recovery use mdns if the cache is unavailable. | 23:16 |
| rjrjr | as it is, the noop cache works fine. it just is slower than using an actual cache. | 23:17 |
| vinod1 | Right now calculating the consensus, gets from mdns and tries to store in the caches to calculate the consensus | 23:17 |
| Kiall | rjrjr: ah, right.. it recovers it rather than prevents it going to ERROR right now.. makes sense | 23:17 |
| paul_glass | rjrjr: I'm seeing zones move to and remain in ERROR indefinitely with the noop cache (with either one or two nameservers) | 23:17 |
| paul_glass | as of latest master | 23:18 |
| rjrjr | that is not how the code i wrote worked. | 23:18 |
| rjrjr | the sync recovery will see it is not in cache and use mdns to get the status. | 23:19 |
| rjrjr | so, you call pool manager to update a zone using noop cache. as kiall explains, there it will go to ERROR state every time because it is looking at cache. but for recovery, if there is no cache, it does the correct thing. | 23:21 |
| paul_glass | rjrjr: I'm seeing these logs when the "Calling periodic_recovery" happens: http://paste.openstack.org/show/185582/ | 23:21 |
| openstackgerrit | Kiall Mac Innes proposed openstack/designate: WIP: Implement TSIG Support in mDNS https://review.openstack.org/160543 | 23:21 |
| Kiall | paul_glass: ahh - looks like the error isn't (directly) cache related: | 23:22 |
| Kiall | u"rndc: 'addzone' failed: already exists\n" | 23:22 |
| Kiall | Could you have leftover zones in bind from a previous install etc of Designate maybe> | 23:22 |
| Kiall | Or does every new zone show the same error message? | 23:23 |
| paul_glass | no, the add zone works the first time. | 23:23 |
| rjrjr | it looks like periodic_recovery is not checking mdns when it does not find the entry in cache. | 23:24 |
| Kiall | Okay, and this only happens if you enable the noop cache? That should be easy enough to reproduce if it happens everytime.. | 23:25 |
| rjrjr | what should happen is periodic_recovery will always check mdns anytime it needs to use cache and cache does not exist. this includes determining whether or not to add the zone. | 23:25 |
| paul_glass | Kiall: yeah, this is only with the noop cache. | 23:26 |
| rjrjr | just a sec. let me look at the patch i submitted. | 23:26 |
| openstackgerrit | Kiall Mac Innes proposed openstack/designate: Implement Scoped TSIG Key Support https://review.openstack.org/160165 | 23:26 |
| openstackgerrit | Kiall Mac Innes proposed openstack/designate: WIP: Implement TSIG Support in mDNS https://review.openstack.org/160543 | 23:26 |
| openstackgerrit | Kiall Mac Innes proposed openstack/designate: Service Class Cleanup - Part 2/3 https://review.openstack.org/160128 | 23:26 |
| Kiall | Last spam attack for tonight, I swear. | 23:26 |
| openstackgerrit | Kiall Mac Innes proposed openstack/designate: Service Class Cleanup - Part 3/3 https://review.openstack.org/160129 | 23:26 |
| openstackgerrit | Kiall Mac Innes proposed openstack/designate: Service Class Cleanup - Part 1/3 https://review.openstack.org/160127 | 23:26 |
| Kiall | Last spam attack for tonight, I swear. ;) | 23:26 |
| vinod1 | rjrjr: _retrieve_from_mdns retrieves from mdns and stores it in the cache - the storing in the cache fails for noop cache | 23:27 |
| rjrjr | vinod1: correct. but for periodic_recovery if there is no cache, i don't use it. | 23:32 |
| *** chlong has joined #openstack-dns | 23:33 | |
| rjrjr | https://review.openstack.org/#/c/149428/11/designate/pool_manager/service.py | 23:34 |
| rjrjr | lines 310 - 340 | 23:34 |
| rjrjr | _retrieve_status() is called with force_check set to True. that means it gets the status from mdns. | 23:35 |
| rjrjr | if the domain exists (and it does in this case) then we log that the domain is already created. | 23:36 |
| vinod1 | If there are multiple servers that need to be updated - how would consensus calculation work without any entries in the cache | 23:36 |
| rjrjr | the same way. we are passing force_check, so it always calls mdns in periodic_recovery if there is not status in cache. | 23:37 |
| rjrjr | _is_create_consensus() is called with force_check set to True. | 23:38 |
| *** openstackgerrit has quit IRC | 23:38 | |
| *** openstackgerrit has joined #openstack-dns | 23:38 | |
| *** ChanServ sets mode: +v openstackgerrit | 23:38 | |
| rjrjr | i wasn't happy with force_check either, but the code did what we wanted it to per the mid cycle meetup. | 23:39 |
| paul_glass | I'm heading out, so I went ahead and submitted a bug for this: https://bugs.launchpad.net/designate/+bug/1427425 | 23:39 |
| openstack | Launchpad bug 1427425 in Designate "Zones remain in ERROR indefinitely with the noop pool manager cache" [Undecided,New] | 23:39 |
| *** paul_glass has quit IRC | 23:40 | |
| rjrjr | https://review.openstack.org/#/c/149428/11/designate/pool_manager/service.py | 23:40 |
| rjrjr | lines 550 - 559 | 23:40 |
| rjrjr | _retrieve_status() will get the status from cache first, if it doesn't exist, it gets it from mdns. | 23:41 |
| openstackgerrit | Graham Hayes proposed openstack/designate: Create ObjectAdaptorMixin and API Mixins https://review.openstack.org/160552 | 23:41 |
| openstackgerrit | Graham Hayes proposed openstack/designate: Move Zones to ObjectAdaptorMixin for API interactions https://review.openstack.org/160553 | 23:41 |
| rjrjr | it looks like you are using check_server the same way, except you don't have all the places that need it covered. | 23:43 |
| vinod1 | I will fix that then. Thanks for looking into this rjrjr | 23:44 |
| rjrjr | just as an aside, a _retrieve_from_cache() that calls mdns seems wrong. that is why i had a _retrieve_status() which called _retrieve_from_cache() and _retrieve_from_mdns() as needed. | 23:45 |
| rjrjr | but that might just be a personal preference. | 23:45 |
| vinod1 | could you explain your reasoning? | 23:46 |
| rjrjr | the method is called _retrieve_from_cache(), but then you have it retrieving it from mdns. | 23:46 |
| vinod1 | ah - now i understand | 23:46 |
| rjrjr | _retrieve_status() doesn't say where you are getting the status. | 23:46 |
| vinod1 | i will check if that would be an easy change to make | 23:47 |
| *** andrewbogott is now known as andrewbogott_afk | 23:57 | |
| *** jmcbride has joined #openstack-dns | 23:57 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!