Monday, 2013-09-16

*** CaptTofu has joined #openstack-dns01:57
*** CaptTofu has quit IRC02:13
*** ekarlso- has quit IRC07:07
openstackgerritKiall Mac Innes proposed a change to stackforge/designate: Add initial BIND9 backend docs  https://review.openstack.org/4670809:44
openstackgerritKiall Mac Innes proposed a change to stackforge/designate: Add initial BIND9 backend docs  https://review.openstack.org/4670809:47
openstackgerritKiall Mac Innes proposed a change to stackforge/designate: Add initial BIND9 backend docs  https://review.openstack.org/4670809:54
openstackgerritA change was merged to stackforge/designate: Add initial BIND9 backend docs  https://review.openstack.org/4670810:48
*** betsy has joined #openstack-dns11:27
*** CaptTofu has joined #openstack-dns11:58
*** CaptTofu has quit IRC12:12
kiallbetsy: welcome to designate-core.. launchpad should have just emailed you (and I have a longer reply en-route ;))12:13
kiallYou'll notice you can now give Code-Review +2, and Approved +1.. or will be able to once Gerrit syncs the changes.12:15
*** CaptTofu has joined #openstack-dns12:23
kiallGERRIT_HOST / GERRIT_PROTO / GERRIT_PORT etc12:27
kiallugh12:27
kiallwrong window12:27
betsykiall: Thanks! Very exciting!13:10
kiallbetsy: yup :)13:13
kiallNow all in need is 10 working fingers and you'll be good to go :D13:14
betsyExactly!13:18
betsyThey're all on the keyboard, but the left hand is definitely still slower13:19
* kiall gets back to actual code again.. 13:21
openstackgerritKiall Mac Innes proposed a change to stackforge/designate: WIP: Introduce the RecordSet concept  https://review.openstack.org/4609413:31
*** eankutse has joined #openstack-dns13:39
openstackgerritKiall Mac Innes proposed a change to stackforge/designate: WIP: Introduce the RecordSet concept  https://review.openstack.org/4609413:51
eankutseKiall: I have the impression that mysqlbind9 backend db has only one table?14:07
*** msisk has joined #openstack-dns14:07
*** betsy has quit IRC14:09
kialleankutse: yea.. I've been wanting to remove that backend.. Since it needs a custom fork of the mysqlbind plugin etc etc14:45
kiallBut.. People seem to like it (or the idea of it anyway ;))14:45
eankutseok. I was asking in connection with the SQLAlchemy improvements to queries for create/update/delete server for the backend.14:48
eankutseIs there a model file for it similar to model file for powerdns?14:48
eankutseShould we invest time in that at this point?14:49
eankutseI mean is it worth investing time in working the create/update/delete server for that backend now regarding https://bugs.launchpad.net/bugs/1214869 ?14:51
kialleankutse: sorry was AFK14:58
kialleankutse: so, honestly.. I've not given it any real thought :) I'd say we should update it like the others..14:58
kiallIt uses SQLAlchemy Core, rather than SQLAlchemy ORM14:59
eankutseok14:59
kiallSo - things are similar but totally different ;014:59
kiall;)*14:59
kiallI've wanted to ditch out (needless) use of the ORM for quite some time.. But it's so low priority ;)15:00
kiallour*15:00
eankutseso is there a schema description to look at?15:00
*** betsy has joined #openstack-dns15:00
kiallHonestly - I don't think I've ever ran the code! Not sure!15:00
kiallMaybe we have that backend raise a NotImplementedException for now15:01
kialland CaptTofu can get a copy of the schema15:01
kiall(CaptTofu is travelling to Seattle today, so I wouldn't expect him to find the time to dig it out over the next few days..)15:02
eankutseok. I'll get the first cut in with the NotImplementedException for now. I'll get with CaptTofu later for more info on the schema.15:03
eankutseExpect a submission by COB today - USA time15:03
kiallAnyway! You reckon the changes for powerdns/bind9 are good?15:03
kiallCool :D15:03
eankutseYes15:03
eankutsepowerdns is good for sure15:04
eankutseBind9 in my opinion would require some re-architecting of the Agent/Pool Manger idea to improve the performance on that end.15:04
kiallYea.. mugsie reckons he'll have the initial pool manager stuff today/tomorrow15:05
eankutseI intend to put in TODO comments in the code to improve that15:05
eankutseright now, there is no really good way to write the zone files but to walk the entire collection15:06
eankutsewe need some thought put into the BInd9 backend updates15:07
kiallYea, I think it was tsimmons who had some ideas around that, using `rndc addzone/rndc delzone` rather write a complete zones.conf15:08
kiallhttps://review.openstack.org/#/c/45745/15:09
eankutsek. But I think as far as this bug fix, we have probably put in as much as we can for now.15:10
CaptTofuhi15:11
eankutseGiven the current state of the Bind9 Backend15:11
eankutseCaptTofu: Hi15:11
kiallYea, I think we ignore the performance issues and fix the issue you're on first, then Tim's code should slot on top to cleanup part of the perf issue..15:11
CaptTofueankutse: I can help you with mysqlbind15:11
kiallAnd .. mugsie pool mgr changes etc15:12
CaptTofuI also really need to talk to the bind people about packaging the code with the other db-backended drivers15:12
eankutseKiall: Yes. I think that's the way to proceed right now.15:13
*** openstackgerrit has quit IRC15:13
eankutseCaptTofu: I was wondering if there is a db schema for the mysqlbind9 backend that I can look at15:14
eankutsefor future reference15:14
*** briancline has quit IRC15:14
kiallv2 recordsets API is coming alive :) http://pastie.org/private/pgvijuencffivh2rit0a15:14
*** briancline has joined #openstack-dns15:17
*** zane has joined #openstack-dns15:24
CaptTofueankutse: yes, indeed with the distribution15:26
CaptTofulet me see if I have it on this laptop15:26
CaptTofueankutse: do you have the git url to the dist?15:26
CaptTofueankutse: https://github.com/CaptTofu/mysql-bind15:27
CaptTofulook in the sql directory15:28
*** openstackgerrit has joined #openstack-dns15:31
*** ChanServ sets mode: +v openstackgerrit15:31
openstackgerritKiall Mac Innes proposed a change to stackforge/designate: WIP: Introduce the RecordSet concept  https://review.openstack.org/4609415:37
*** eankutse has quit IRC15:40
betsykiall: Yes. I've been following your WIP on the conversion to RecordSets15:42
kiallbetsy: not sure that's possible.. Diff is *wayy* too big15:42
kiall;)15:42
betsyYeah, but not really not many changes per file. Doesn't take long to skim thru it and follow along, especially since I was very familiar with all the record code after doing the documentation for it15:43
kiall:)15:44
betsyNice to see some actual output from it15:44
kiallJust got that in the last few mins ;)15:44
betsyCool15:44
kiallJust hacking in something to format the records array correctly for A/AAAA/MX records.. ignoring the rest as I want to make something extensible there!15:45
betsyI was assuming that you could have multiple A record sets as long as they had different names, and your example verifies that assumption15:46
kiallYea, each RRSset is unique on (name, class, type) .. and we don't support CH records .. only IN, so class is essentially ignored for us15:47
*** eankutse has joined #openstack-dns15:48
kiallhttp://pastie.org/private/psjcabcaqrmrygvds6jqw15:51
kiallOutputs from a few of the API calls15:51
eankutseCaptTofu: Thanks, I'll check that out.15:51
betsykiall: And zones!15:52
kiallbetsy: ?16:20
betsyYep16:20
kiallNot sure what you meant "kiall: And zones!"16:20
betsyJust looking at the other pastie you sent out16:21
betsyIt's got recordsets and zones. I hadn't seen the name change from domain to zone16:21
betsyNothing significant16:22
betsy:)16:22
betsyv2 looks like it's starting to come together16:22
kiallah :) Yea, they are pretty similar :)16:22
*** CaptTofu has quit IRC16:36
openstackgerritKiall Mac Innes proposed a change to stackforge/designate: WIP: Introduce the RecordSet concept  https://review.openstack.org/4609416:39
*** ekarlso has joined #openstack-dns16:49
*** vipul is now known as vipul-away17:16
*** vipul-away is now known as vipul17:22
*** zane has quit IRC17:26
*** jmcbride has joined #openstack-dns17:43
*** jmcbride has quit IRC18:04
*** eankutse has quit IRC18:04
*** jmcbride has joined #openstack-dns18:09
*** zane has joined #openstack-dns18:13
*** eankutse has joined #openstack-dns18:15
*** eankutse has quit IRC18:15
*** eankutse has joined #openstack-dns18:15
*** jmcbride has quit IRC18:28
*** jmcbride1 has joined #openstack-dns18:29
openstackgerritKiall Mac Innes proposed a change to stackforge/designate: WIP: Introduce the RecordSet concept  https://review.openstack.org/4609418:32
*** CaptTofu has joined #openstack-dns18:52
*** artom has joined #openstack-dns18:55
artomHi again. Seeking confirmation that what I understood by looking at the code is correct.18:55
artomI need a en external script to be called when a domain is deleted.18:56
artomWhile hardcoding it into one of the many delete_domain methods (probably Central's) would be the first reflex, a more "well-designed" solution would be an AMQP client that receives the domain.deleted notifications that are sent out by Central's delete_domain.18:57
artom(Not too familiar with AMQP, so please forgive any wrong vocabulary).18:58
*** zane has quit IRC20:12
*** zane has joined #openstack-dns20:12
*** vinodmr has joined #openstack-dns20:15
openstackgerritBetsy Luzader proposed a change to stackforge/designate: Give more meaningful error message when no servers are configured  https://review.openstack.org/4680920:41
*** shakayumi has joined #openstack-dns20:42
*** vipul is now known as vipul-away21:00
*** vipul-away is now known as vipul21:15
*** vinodmr has quit IRC21:15
*** eankutse has quit IRC21:36
*** jmcbride1 has quit IRC22:07
*** shakayumi has quit IRC22:08
*** betsy has quit IRC22:09
*** shakayumi has joined #openstack-dns22:09
*** eankutse has joined #openstack-dns22:29
*** eankutse has quit IRC22:29
*** eankutse has joined #openstack-dns22:30
*** eankutse has quit IRC22:55
*** eankutse has joined #openstack-dns23:00
*** zane has quit IRC23:01
*** msisk has quit IRC23:19
*** eankutse has quit IRC23:22
*** eankutse has joined #openstack-dns23:39

Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!