| *** rmoe has quit IRC | 00:14 | |
| *** eankutse has joined #openstack-dns | 00:14 | |
| *** eankutse has quit IRC | 00:17 | |
| *** eankutse has joined #openstack-dns | 00:17 | |
| *** darshan104 has quit IRC | 00:30 | |
| *** rmoe has joined #openstack-dns | 00:35 | |
| *** russellb has quit IRC | 00:46 | |
| *** russellb has joined #openstack-dns | 00:49 | |
| *** nosnos has joined #openstack-dns | 01:54 | |
| *** amcrn has quit IRC | 01:55 | |
| *** eankutse1 has joined #openstack-dns | 02:23 | |
| *** eankutse has quit IRC | 02:26 | |
| *** eankutse1 has quit IRC | 02:27 | |
| *** richm has quit IRC | 02:29 | |
| *** k4n0 has joined #openstack-dns | 05:45 | |
| *** darshan104 has joined #openstack-dns | 06:09 | |
| *** k4n0 has quit IRC | 07:07 | |
| *** k4n0 has joined #openstack-dns | 07:22 | |
| openstackgerrit | OpenStack Proposal Bot proposed a change to openstack/designate: Updated from global requirements https://review.openstack.org/120331 | 07:33 |
|---|---|---|
| *** darshan104 has quit IRC | 08:10 | |
| *** nosnos has quit IRC | 09:41 | |
| *** nosnos has joined #openstack-dns | 09:42 | |
| *** nosnos has quit IRC | 09:47 | |
| openstackgerrit | Jordan Pittier proposed a change to openstack/designate: Devstack setup : use get_or_create_* helpers for user and endpoint creation https://review.openstack.org/120363 | 10:02 |
| openstackgerrit | A change was merged to openstack/designate: Add index on powerdns.records.designate_record_id https://review.openstack.org/120086 | 10:53 |
| openstackgerrit | Jordan Pittier proposed a change to openstack/designate: Devstack setup : use get_or_create_* helpers https://review.openstack.org/120363 | 11:19 |
| *** ekarlso- has quit IRC | 11:37 | |
| *** ekarlso- has joined #openstack-dns | 11:37 | |
| *** mwagner_lap has quit IRC | 11:38 | |
| openstackgerrit | A change was merged to openstack/designate: Updated from global requirements https://review.openstack.org/120331 | 12:06 |
| openstackgerrit | Jordan Pittier proposed a change to openstack/designate: Devstack Integration : Fix Bashate E20 https://review.openstack.org/120399 | 12:21 |
| *** mwagner_lap has joined #openstack-dns | 12:37 | |
| *** mwagner_lap has quit IRC | 12:41 | |
| *** mwagner_lap has joined #openstack-dns | 12:41 | |
| *** jordanP has joined #openstack-dns | 12:43 | |
| Kiall | jordanP: I'm guessing you're the Jordan Pittier submitting some devstack fixes? | 12:56 |
| jordanP | Kiall, that's me sir :) | 12:56 |
| Kiall | Cool - Thanks :) | 12:57 |
| jordanP | those are not real issues but I saw somewhere that designate could be integrated into devstack soon | 12:57 |
| Kiall | I'm guessing you're more familiar with devstack than I, do you happen to know if we have any other issues etc we should fix before we try and have it merged? | 12:57 |
| jordanP | yeah, there the use of local variable too | 12:57 |
| jordanP | they should be lower case | 12:58 |
| jordanP | dean troyer made a cleanup series of patch recently, it will be nice to follow the "convention", though there's not a bashate check for this yet | 12:58 |
| jordanP | I mean, in lib/designate all local variables should be in lower case | 12:59 |
| Kiall | So - Style stuff, that's almost always an easy fix :) Glad there's no bigger issues (that we know of ;)) | 12:59 |
| jordanP | yeah, Ive been playing with designate in a devstack, so far so good | 12:59 |
| Kiall | Cool :) | 13:01 |
| *** richm1 has joined #openstack-dns | 13:07 | |
| openstackgerrit | A change was merged to openstack/designate: Devstack setup : use get_or_create_* helpers https://review.openstack.org/120363 | 13:10 |
| *** nkinder has quit IRC | 13:13 | |
| *** eankutse has joined #openstack-dns | 13:13 | |
| *** eankutse has quit IRC | 13:18 | |
| *** eankutse has joined #openstack-dns | 13:18 | |
| *** eankutse has quit IRC | 13:25 | |
| openstackgerrit | Kiall Mac Innes proposed a change to openstack/designate: Backends need to implement c/u/d_recordset https://review.openstack.org/119768 | 13:26 |
| *** jmcbride has joined #openstack-dns | 13:32 | |
| *** rmoe has quit IRC | 13:35 | |
| *** rmoe has joined #openstack-dns | 13:37 | |
| openstackgerrit | Jordan Pittier proposed a change to openstack/designate: Devstack Integration : Coding style fix https://review.openstack.org/120399 | 13:38 |
| *** eankutse has joined #openstack-dns | 13:45 | |
| *** jmcbride has quit IRC | 13:47 | |
| *** jmcbride has joined #openstack-dns | 13:58 | |
| *** timsim has joined #openstack-dns | 14:00 | |
| *** nkinder has joined #openstack-dns | 14:03 | |
| *** eankutse has quit IRC | 14:04 | |
| *** eankutse has joined #openstack-dns | 14:07 | |
| *** zain has joined #openstack-dns | 14:32 | |
| *** zain has joined #openstack-dns | 14:32 | |
| *** EricGonczer_ has joined #openstack-dns | 14:33 | |
| *** paul_glass has joined #openstack-dns | 14:49 | |
| jordanP | mugsie, my patch 'Devstack Integration : Coding style fix' failed, sorry about that, i missed something | 14:49 |
| openstackgerrit | Jordan Pittier proposed a change to openstack/designate: Devstack Integration : Coding style fix https://review.openstack.org/120399 | 14:56 |
| *** EricGonczer_ has quit IRC | 14:58 | |
| *** EricGonczer_ has joined #openstack-dns | 14:59 | |
| *** EricGonczer_ has quit IRC | 15:01 | |
| *** jmcbride has quit IRC | 15:01 | |
| *** EricGonczer_ has joined #openstack-dns | 15:01 | |
| *** EricGonc_ has joined #openstack-dns | 15:02 | |
| *** eankutse has quit IRC | 15:06 | |
| *** eankutse1 has joined #openstack-dns | 15:06 | |
| *** EricGonczer_ has quit IRC | 15:06 | |
| *** paul_glass1 has joined #openstack-dns | 15:08 | |
| *** paul_glass has quit IRC | 15:11 | |
| *** vinod1 has joined #openstack-dns | 15:28 | |
| *** jmcbride has joined #openstack-dns | 15:32 | |
| *** k4n0 has quit IRC | 15:34 | |
| *** jmcbride has quit IRC | 15:40 | |
| *** diga has joined #openstack-dns | 15:42 | |
| *** eankutse1 has quit IRC | 15:58 | |
| jordanP | guys, I can't find a designate client for api v2, is that normal ? | 15:59 |
| *** eankutse has joined #openstack-dns | 15:59 | |
| vinod1 | jordanP - yes currently the designate client only supports api v1 | 16:00 |
| jordanP | vinod1, is there ongoing work to provide a client for api v2 ? | 16:00 |
| timsim | There was some work done, but it never merged: https://review.openstack.org/#/q/project:openstack/designateclient,n,z | 16:00 |
| jordanP | cause I love the abilitity to set a ptrrecord on a floating ip | 16:00 |
| jordanP | timsim, you meant https://review.openstack.org/#/q/project:openstack/python-designateclient,n,z :) | 16:02 |
| timsim | For some reason this link did not copy https://review.openstack.org/#/c/76457/ | 16:02 |
| jordanP | timsim, do you know why https://review.openstack.org/#/c/76457/ was abandonned ? | 16:03 |
| vinod1 | ekarlso- might know better | 16:04 |
| jordanP | i have to leave, i'll ask tomorrow | 16:07 |
| *** jmcbride has joined #openstack-dns | 16:07 | |
| *** jordanP has quit IRC | 16:08 | |
| *** mwagner_lap has quit IRC | 16:09 | |
| *** rmoe has quit IRC | 16:13 | |
| Kiall | vinod1: it was because the v2 api wasn't final :) | 16:22 |
| vinod1 | that is what i assumed - but I wasn't sure - jordanP the originator of the question isn't around now | 16:24 |
| ekarlso- | Kiall: regarding that... | 16:24 |
| ekarlso- | I got branches locally ^ | 16:24 |
| *** rmoe has joined #openstack-dns | 16:30 | |
| *** jmcbride has quit IRC | 16:51 | |
| *** rjrjr has joined #openstack-dns | 16:57 | |
| *** rossk has joined #openstack-dns | 17:05 | |
| *** amcrn has joined #openstack-dns | 17:05 | |
| *** richm1 has quit IRC | 17:20 | |
| *** mwagner_lap has joined #openstack-dns | 17:33 | |
| *** richm1 has joined #openstack-dns | 17:34 | |
| vinod1 | rjrjr: I did read the specs | 17:37 |
| rjrjr | i have a 'status' in the database table for pool manager, but i only need that for create/delete of a zone. | 17:37 |
| rjrjr | the way i envision pool manager working for an update zone is: | 17:38 |
| rjrjr | loop through servers and call mini-dns zone change method | 17:39 |
| rjrjr | loop through servers and call poll for serial number | 17:39 |
| rjrjr | mini-dns' poll for serial number method will only report back to pool manager when it is successful (need context, domain, and serial number) | 17:40 |
| rjrjr | update_status (the call that mini-dns makes) updates the database and checks if it is has enough info to call update_status in central | 17:40 |
| rjrjr | it will base that call to central on whether or not the threshold number of servers has a new (i.e. higher) serial number than what the Designate database knows about | 17:41 |
| rjrjr | then the update_status in central updates the Designate records table and changes pendings to active or active to deleted based on serial number | 17:42 |
| vinod1 | so would the status be just pending and never move to error? | 17:44 |
| vinod1 | what if the threshold is 100% and one of the name servers is down? | 17:44 |
| rjrjr | i don't know how you would move an update of an individual record to an error state. | 17:44 |
| rjrjr | what we could do is look at the pool manager's status table and report an error if a server is lagging the others or some other way of getting an error. | 17:46 |
| vinod1 | that is what i was thinking | 17:46 |
| rjrjr | but we can't base it on a particular call the user makes, otherwise we would need to track the through pool manager, which would balloon the pool manager status table. | 17:47 |
| vinod1 | if we track the time in the pool manager status table too then periodically a thread could see something was pending for a long time - then send it as error to central | 17:47 |
| rjrjr | that is the only way i see that working without impacting performance negatively. | 17:48 |
| rjrjr | but figuring out which records are in 'Error' is going to be tricky. | 17:48 |
| rjrjr | do you mark all records with a serial number or higher as Error? or just some of them? | 17:49 |
| vinod1 | once pool manager calls central's update_status with a serial number then | 17:49 |
| vinod1 | central would mark all pending recordsets/domain changes upto that serial as error | 17:49 |
| rjrjr | remember, we are recording the 'desired' serial number when a request to update the zone comes in. | 17:50 |
| rjrjr | how do we know the server is down versus just being slow? | 17:50 |
| rjrjr | does it even matter? | 17:50 |
| vinod1 | the time should be configurable based on the admin's preferences | 17:51 |
| rjrjr | i can do that. | 17:51 |
| rjrjr | i'll add this. | 17:51 |
| rjrjr | again, we don't need mini-dns to report the status to pool manager. i don't even need to know about the error'd requests. | 17:52 |
| rjrjr | that will help cut down on calls between mini-dns and pool manager. | 17:52 |
| rjrjr | (and hence calls to the pool manager table.) | 17:52 |
| vinod1 | how will pool manager know when a particular change is on the nameserver? | 17:53 |
| rjrjr | mini-dns is sending successful polls back to pool manager. | 17:53 |
| rjrjr | which should be most polls anyway (unless there is a problem.) | 17:54 |
| rjrjr | which we'll report as you suggested. | 17:54 |
| vinod1 | Essentially you are saying that minidns will call update_status only on a success? | 17:55 |
| rjrjr | yes. | 17:55 |
| rjrjr | we are giving minidns X amount of time to get a success. (defaults to about 1.5 minutes.) | 17:55 |
| vinod1 | the minidns has retries and expire based on the number of retries - whereas the pool manager would mark something as error based on time | 17:55 |
| rjrjr | time or attempts or number of serial numbers we are behind or ... | 17:56 |
| rjrjr | we can base it on any number of things. | 17:56 |
| vinod1 | so if minidns has 3 retries and it takes 3 minutes - minidns could still be retrying when pool manager marks something as error | 17:56 |
| rjrjr | yes. | 17:57 |
| vinod1 | so we need to handle the case when minidns reports a success for something that the pool manager already marked as an error | 17:57 |
| rjrjr | that is built in already. | 17:58 |
| vinod1 | Ideally if pool manager is basing everything on time, then minidns too should do the same | 17:58 |
| rjrjr | vinod: think of the case where someone wants to add 100 records and writes a program or script to do that. think of how many calls we are talking about if we had 10 servers in 1 pool. | 17:58 |
| rjrjr | Designate doesn't have a 'batch' call to add all 100 at once. | 17:59 |
| rjrjr | we are looking at 1000 calls to mini-dns being fired off pretty quickly. | 17:59 |
| vinod1 | we did talk about adding 'bulk' calls being added to designate | 18:00 |
| rjrjr | sure, but we don't have it right now. and it doesn't stop a user from doing it one record at a time even if we did. | 18:00 |
| vinod1 | in this particular case, we could have the pool manager do an optimization - delay sending requests to minidns for sometime, if an update comes in between then repeat the process | 18:01 |
| rjrjr | an atomic call is going to be needed if we want to support nsupdate (dynamic DNS) regardless. | 18:01 |
| rjrjr | that is too much. | 18:01 |
| rjrjr | or we just make sure the calls between pool manager and mini-dns are pertinent and only for successes. | 18:01 |
| *** diga has quit IRC | 18:01 | |
| rjrjr | that is how i designated the pool manage update_status call. | 18:02 |
| vinod1 | How many errors would we normally see - would that save much - or would it just complicate the design? | 18:03 |
| rjrjr | it saves a call and talking to the database. | 18:03 |
| rjrjr | i'll just ignore errors regardless in pool manager. 8^) | 18:04 |
| rjrjr | IMO, chatting on rabbitmq should be kept minimal and calling the database from pool manager should be minimum. | 18:04 |
| rjrjr | the design is simpler by not passing a status. | 18:05 |
| vinod1 | if you don't update the database, on error, how would the pool manager keep track of errors later if minidns reports a success? | 18:05 |
| rjrjr | i'm not tracking errors, i'm tracking serial numbers. | 18:05 |
| rjrjr | errors shouldn't matter. | 18:05 |
| rjrjr | in order for me to track error, i would need to know for what, then i'm back to having a request ID and the database grows exponentially. | 18:06 |
| vinod1 | i think we are optimizing prematurely even before we have something working correctly | 18:06 |
| openstackgerrit | Doug Hellmann proposed a change to openstack/designate-specs: Add RSS feed https://review.openstack.org/120501 | 18:07 |
| rjrjr | vinod, one of the things that was made clear to me in the offsite is the pool manager database cannot have a row for every request. | 18:07 |
| openstackgerrit | Doug Hellmann proposed a change to openstack/designate-specs: Remove templates from toctrees https://review.openstack.org/120502 | 18:07 |
| openstackgerrit | Doug Hellmann proposed a change to openstack/designate-specs: Use the current date for the copyright statement https://review.openstack.org/120503 | 18:07 |
| rjrjr | we have a row for every server - zone update | 18:07 |
| rjrjr | potentially a row for every server zone create | 18:08 |
| rjrjr | potentially a row for every server - zone delete | 18:08 |
| rjrjr | if i have 10 servers in 1 pool that is 30 rows potentially | 18:08 |
| rjrjr | trust me when i'm telling you, i followed KISS when i did this design. | 18:09 |
| rjrjr | i kept it as simple as i possibly could. | 18:09 |
| rjrjr | optimization wasn't why it turned out this way, it was to keep it simple. | 18:10 |
| vinod1 | did we talk about having the pool_manager_status in memcache? | 18:11 |
| rjrjr | we did. everyone wanted to try database first. it doesn't fix the issue either. | 18:11 |
| rjrjr | if i track a single request through pool manager, we have to track all 10 servers. what if one is down. do i hold onto those requests individually for hours or days until that server is back online? | 18:12 |
| rjrjr | or do i remove them when the threshold is met? how do i know when a call to update_status is for a request that has "already been handled" and removed or for a new request? you can see how complicated this gets. | 18:13 |
| rjrjr | it seems to me what is important is the zone and the serial number. | 18:14 |
| rjrjr | so, the status table stores the server - zone - serial number | 18:14 |
| rjrjr | (for updates) | 18:14 |
| rjrjr | one row regardless of how many requests. i just keep updating the serial number and once the lowest serial number for a threshold increases, that is time to call central's update_status. | 18:15 |
| rjrjr | simple and elegant. | 18:15 |
| rjrjr | i do store status for create and delete in the same table. | 18:16 |
| *** vinod1 has quit IRC | 18:21 | |
| *** vinod1 has joined #openstack-dns | 19:11 | |
| *** mgagne has quit IRC | 19:18 | |
| *** mgagne has joined #openstack-dns | 19:19 | |
| *** mgagne is now known as Guest75250 | 19:19 | |
| *** rossk has quit IRC | 19:25 | |
| openstackgerrit | Doug Hellmann proposed a change to openstack/designate-specs: Remove templates from toctrees https://review.openstack.org/120502 | 19:40 |
| openstackgerrit | Doug Hellmann proposed a change to openstack/designate-specs: Use the current date for the copyright statement https://review.openstack.org/120503 | 19:40 |
| *** sballe has joined #openstack-dns | 19:52 | |
| *** vinod1 has quit IRC | 19:59 | |
| *** tmc3inphilly has joined #openstack-dns | 20:00 | |
| *** tmc3inphilly has left #openstack-dns | 20:00 | |
| openstackgerrit | Doug Hellmann proposed a change to openstack/designate-specs: Remove templates from toctrees https://review.openstack.org/120502 | 20:01 |
| openstackgerrit | Doug Hellmann proposed a change to openstack/designate-specs: Use the current date for the copyright statement https://review.openstack.org/120503 | 20:01 |
| openstackgerrit | Doug Hellmann proposed a change to openstack/designate-specs: Add RSS feed https://review.openstack.org/120501 | 20:01 |
| *** vinod1 has joined #openstack-dns | 20:06 | |
| *** jmcbride has joined #openstack-dns | 20:14 | |
| *** vinod1 has quit IRC | 20:18 | |
| *** vinod1 has joined #openstack-dns | 20:20 | |
| *** jmcbride has quit IRC | 20:21 | |
| *** vinod1 has quit IRC | 20:31 | |
| *** amcrn has quit IRC | 20:35 | |
| *** vinod1 has joined #openstack-dns | 20:44 | |
| *** jmcbride has joined #openstack-dns | 21:05 | |
| *** paul_glass has joined #openstack-dns | 21:11 | |
| *** paul_glass1 has quit IRC | 21:11 | |
| *** vinod1 has quit IRC | 21:17 | |
| *** paul_glass has quit IRC | 21:18 | |
| *** vinod1 has joined #openstack-dns | 21:19 | |
| *** EricGonc_ has quit IRC | 21:23 | |
| *** jmcbride has quit IRC | 21:27 | |
| *** jmcbride has joined #openstack-dns | 21:29 | |
| *** vinod1 has quit IRC | 21:31 | |
| *** vinod1 has joined #openstack-dns | 21:42 | |
| *** sballe_ has joined #openstack-dns | 21:46 | |
| *** timsim has quit IRC | 21:47 | |
| *** jmcbride has quit IRC | 21:49 | |
| *** sballe has quit IRC | 21:49 | |
| *** amcrn has joined #openstack-dns | 21:53 | |
| *** eankutse has quit IRC | 22:07 | |
| *** nkinder has quit IRC | 22:12 | |
| *** zain has quit IRC | 22:35 | |
| *** nkinder has joined #openstack-dns | 23:02 | |
| openstackgerrit | OpenStack Proposal Bot proposed a change to openstack/python-designateclient: Updated from global requirements https://review.openstack.org/120608 | 23:09 |
| *** EricGonczer_ has joined #openstack-dns | 23:20 | |
| openstackgerrit | Vinod Mangalpally proposed a change to openstack/designate: Getting deleted recordsets returns a 404 https://review.openstack.org/120619 | 23:33 |
| *** vinod1 has quit IRC | 23:34 | |
| *** richm1 has quit IRC | 23:44 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!