Wednesday, 2014-11-26

*** nkinder has joined #openstack-dns00:52
*** rmoe has quit IRC01:00
*** rmoe has joined #openstack-dns01:17
*** nosnos has joined #openstack-dns02:14
*** richm has quit IRC02:27
*** vinod has joined #openstack-dns02:28
*** vinod has quit IRC02:31
*** vipul has quit IRC02:33
*** vipul has joined #openstack-dns02:37
*** richm has joined #openstack-dns02:44
*** richm has quit IRC02:52
*** nosnos has quit IRC03:01
*** ryanpetrello has joined #openstack-dns03:18
*** ryanpetrello has quit IRC03:41
*** amcrn has quit IRC03:47
*** ryanpetrello has joined #openstack-dns04:07
*** nosnos has joined #openstack-dns04:16
*** ryanpetrello has quit IRC04:23
*** iraw_ has joined #openstack-dns04:40
*** ryanpetrello has joined #openstack-dns04:49
*** iraw_ has quit IRC05:13
*** k4n0 has joined #openstack-dns06:03
*** ryanpetrello has quit IRC06:05
*** nihilifer has joined #openstack-dns07:05
*** yingjun has joined #openstack-dns07:47
yingjunHi, anyone there? i got an error when running devstack with designate, could anyone help?07:48
yingjunNoSuchMethod: Endpoint does not support RPC method create_server designate07:49
*** ryanpetrello has joined #openstack-dns07:51
*** ryanpetrello has quit IRC07:56
*** GonZo2K has joined #openstack-dns08:12
*** yingjun has quit IRC08:41
*** jordanP has joined #openstack-dns08:54
*** yingjun has joined #openstack-dns09:08
*** GonZo2K has quit IRC09:38
*** rjrjr has joined #openstack-dns10:03
rjrjris anyone else having problems running the 'tox -e py27 -r' with the latest code?10:04
rjrjri'm getting failures in the zone transfer code.10:04
rjrjrHere is the output of the tox on a brand new 'git clone' - http://paste.openstack.org/show/138731/10:10
*** jmcbride has joined #openstack-dns10:22
*** rediskin has joined #openstack-dns10:31
*** jmcbride has quit IRC10:33
*** jmcbride has joined #openstack-dns10:33
rjrjrin addition to the unit tests failing, when i'm running central, i'm now getting errors concerning the zone transfer request - http://paste.openstack.org/show/138748/10:45
ekarlso-rjrjr: ?11:11
*** yingjun has joined #openstack-dns11:13
yingjunHelp! i got an error when creating server with command: designate server-create --name ns1.devstack.org. Backend: Unknown backend failure: RemoteError(u'Remote error: NoSuchMethod Endpoint does not support RPC method create_server11:16
yingjunanyone know what’s happened ?11:16
yingjun?11:16
rjrjrekarlso, i'm here.11:19
*** jordanP has quit IRC11:23
rjrjri'm block right now on my code.  if anyone can help me figure out why the transfer zone code is not working, it would be greatly appreciated. 8^)11:55
rjrjrignore me.  not sure why, but tests are working now.12:29
*** jordanP has joined #openstack-dns13:06
*** jordanP has quit IRC13:10
*** jordanP has joined #openstack-dns13:18
*** nosnos has quit IRC13:20
openstackgerritRon Rickard proposed openstack/designate: Pool Manager  https://review.openstack.org/13165613:26
*** ryanpetrello has joined #openstack-dns13:42
*** yingjun has quit IRC13:43
*** iraw_ has joined #openstack-dns14:03
*** iraw_ has quit IRC14:03
*** nkinder has quit IRC14:10
*** RaginBajin has joined #openstack-dns14:11
*** timsim has joined #openstack-dns14:12
openstackgerritEndre Karlson proposed openstack/designate: Support secondary zones  https://review.openstack.org/13368214:22
openstackgerritEndre Karlson proposed openstack/designate: MDNS part of Secondary zones  https://review.openstack.org/13368314:22
*** vinod has joined #openstack-dns14:31
*** mikeit has joined #openstack-dns14:42
*** mikeit has quit IRC14:47
*** mikeit has joined #openstack-dns14:48
*** rediskin has left #openstack-dns14:54
rjrjrvinod: you around?14:54
ekarlso-rjrjr: got your stuff going ?14:56
rjrjryes.14:57
rjrjrekarlso: do you know how i add items to today's meeting agenda?14:57
rjrjri always seem to update the wrong section on the meeting page.14:58
*** nkinder has joined #openstack-dns14:58
rjrjrdo i add to the future/unscheduled agenda items?  agenda?14:59
ekarlso-rjrjr: not sure :|15:00
*** paul_glass has joined #openstack-dns15:00
rjrjri've been in this place before.  i'm updating both. 8^)15:00
*** vinod has quit IRC15:01
*** jmcbride has quit IRC15:04
*** jmcbride has joined #openstack-dns15:06
*** betsy has quit IRC15:12
*** nihilifer has quit IRC15:23
*** vinod has joined #openstack-dns15:31
vinodrjrjr: I am here15:31
vinodLooking at your changes now15:31
rjrjrvinod: we have 2 issues and I'm looking for advice on how to fix one of them.15:34
rjrjrwe are calling notify_zone_changed followed by poll_for_serial_number.  the problem, there is not a sufficient delay between the call.  the result, the poll_for_serial_number fails to get the newest serial number.15:35
rjrjrso, we obviously need a delay, just wanting to hear ideas on where that delay should be.15:36
rjrjrsince mdns has the other timing parameters (retry, interval, timeout) it makes sense to me to introduce a 'poll_delay' parameter there..15:36
vinodi was expecting to see this issue15:37
rjrjri would set this to 0 for notify_zone_changed and to the value in the config file (i'll add poll_delay to [service:pool_manager].15:37
rjrjrthe config value for poll_for_serial_number.  thoughts?15:38
rjrjri'll also set the default to 1 second.15:39
vinodso would pool manager wait for the delay ( of 1 second) before issuing a poll_for_serial_number?15:42
rjrjrno, i was thinking about adding this to mdns.15:42
rjrjrjust to be consistent since all the other timing parameters are being passed to mdns.15:43
rjrjrit would be another parameter to poll_for_serial_number (and even notify_zone_changed).15:44
rjrjri would set it to 0 from pool manager for notify_zone_changed.  i would set it to config file value for poll_for_serial_number.15:44
rjrjrmy thinking, it is just another timing parameter, so it should be passed to mdns like the others.15:45
vinodcurrently poll_for_serial_number has a retry logic - i was wondering if there is a bug there15:46
rjrjrbug in the retry logic?15:46
rjrjryeah, probably.15:46
rjrjrthe retry_interval isn't being honored for retries that are not timeouts.15:46
vinodwe should fix that so that it always sleeps for retry_interval between retries15:48
vinodthat would i think fix the problem15:49
*** GonZo2K has joined #openstack-dns15:49
rjrjrno, because the first attempt to get the serial number will always fail and trigger another attempt.15:49
rjrjri personally think that is a sloppy way of handling the issue.15:49
rjrjrhence my wanting to introduce the delay.15:50
rjrjr(also a 1 second delay is preferrable to a 2 second retry interval.)15:50
*** betsy has joined #openstack-dns15:50
vinodthat is true15:53
rjrjrokay, i'll introduce the delay parameter.  if reviewers don't like it, they can speak up in the review stage. 8^)15:54
vinodanother alternative to delays is to communicate the status of notify_zone_changed or for mdns to call poll_for_serial_number after a successful call to notify_zone_changed15:54
rjrjrtrue.  but, the code is mostly working.  this delay thing is one of a few items that needs to be hammered out.15:56
*** EricGonczer_ has joined #openstack-dns15:56
*** paul_glass has quit IRC15:56
*** GonZo2K has quit IRC15:57
rjrjrthe only other bug I'm seeing has to do with NS and SOA records in the records table.  right now, they have a serial number of 1 and the status/action is wrong initially.15:57
rjrjri know where i need to fix that.  other than the delay and that issue, the code is working.15:57
*** paul_glass has joined #openstack-dns15:57
rjrjri was going to fire up a few DNS servers today to do more testing.  and i still have lots of unit tests to write.15:58
rjrjrbut, i'm hoping none of these will hold back getting this code reviewed.15:58
*** paul_glass has quit IRC15:58
*** jmcbride has quit IRC15:58
*** jmcbride has joined #openstack-dns15:58
rjrjrdid find a bug with the ZoneTransferRequest object, but that is a separate issue entirely. 8^)15:59
rjrjr(the object extends PersistentObjectMixin and also defines the id, created_at, update_at, and versions attributes.)16:00
*** EricGonczer_ has quit IRC16:01
vinodI saw some issue at startup with periodic-sync16:01
vinodthat should be a minor issue if at all16:02
rjrjrhmmm... okay, i haven't seen any issue with that, but i'm pretty close to the code, so maybe i overlooked something.16:03
*** mikeit has quit IRC16:04
vinodso are you planning to fix mdns too for the retry_interval or do you want me to  fix it?16:06
*** ryanpetrello has quit IRC16:06
rjrjri'll take care of it.16:08
rjrjrtaking a break until meeting.  i'll fix all this after the meeting.  shouldn't take but an hour.16:09
rjrjralso need to use passed parameters in notify_zone_changed.16:13
vinodwhat do you mean?16:14
*** zigo has quit IRC16:16
*** EricGonczer_ has joined #openstack-dns16:18
*** EricGonczer_ has quit IRC16:23
vinodrjrjr: This is the error in the pool manager log related to periodic sync that I was talking about - http://paste.openstack.org/show/KXiLsL7j4WhfpvC0tJca/16:24
rjrjri'm not seeing it.  are you doing anything to induce the error?16:26
*** zigo has joined #openstack-dns16:27
*** paul_glass has joined #openstack-dns16:27
*** k4n0 has quit IRC16:30
vinodno - this is just after starting the pool manager service16:31
*** ryanpetrello has joined #openstack-dns16:31
rjrjris all your code up to date (i.e. if you are using devstack, did you let it get the latest code including oslo?)16:35
rjrjri believe the olso library i'm pulling that from was updated a few weeks back.16:35
*** ryanpetrello_ has joined #openstack-dns16:45
*** ryanpetrello has quit IRC16:48
*** ryanpetrello_ is now known as ryanpetrello16:48
*** rjrjr has quit IRC16:56
*** rjrjr_ has joined #openstack-dns16:58
*** ryanpetrello has quit IRC17:12
*** rediskin has joined #openstack-dns17:27
*** paul_glass has quit IRC17:35
*** ryanpetrello has joined #openstack-dns17:35
*** johnbelamaric has joined #openstack-dns17:36
*** rmoe has quit IRC17:40
*** rjrjr_ has quit IRC17:46
*** rjrjr has joined #openstack-dns17:46
*** rmoe has joined #openstack-dns18:00
*** ryanpetrello_ has joined #openstack-dns18:01
*** jordanP has quit IRC18:04
*** ryanpetrello has quit IRC18:04
*** ryanpetrello_ is now known as ryanpetrello18:04
*** rjrjr has quit IRC18:11
timsimFYI: Most of the Americans are on vacation for the rest of the week :) Tomorrow is Thanksgiving, the one day a year where we give thanks and eat a lot with people who annoy us ;)18:28
mugsietimsim: :)18:30
*** ryanpetrello_ has joined #openstack-dns18:30
*** betsy has quit IRC18:31
*** ryanpetrello has quit IRC18:33
*** ryanpetrello_ is now known as ryanpetrello18:33
Kialltimsim: lol18:41
KiallIt's also the one day of the year I have zero meetings :P18:41
ahucongrats18:51
ekarlso-any labrats for the v2 client ?19:11
*** timsim has quit IRC19:15
*** johnbelamaric has left #openstack-dns19:22
*** bitblt has joined #openstack-dns19:28
*** bitblt has left #openstack-dns19:30
*** GonZo2K has joined #openstack-dns19:56
*** ToMiles has joined #openstack-dns19:56
*** ryanpetrello has quit IRC20:00
*** ryanpetrello has joined #openstack-dns20:02
*** ryanpetrello has quit IRC20:10
*** ryanpetrello has joined #openstack-dns20:20
*** betsy has joined #openstack-dns20:25
*** nkinder has quit IRC20:28
*** ryanpetrello has quit IRC21:32
*** rjrjr has joined #openstack-dns21:40
rjrjrkiall: you still around?21:41
rjrjractually, anyone who is familiar with the storage code around?21:42
Kiallheya21:43
*** nkinder has joined #openstack-dns21:45
Kiallrjrjr: ^21:46
rjrjri think i won't be able to do what i want via storage after all.  i need to not include records that are about to be deleted (record.action = 'DELETE') from the AXFR.21:49
rjrjri think i'm going to have to do that in the AXFR code.21:50
rjrjrso, ignore me. 8^)21:50
KiallCan we not do find_recordsets(domain_id, criterion={"action": "!DELETE"})21:51
rjrjrrecordset doesn't have action.  that is on a record.21:52
rjrjrsorry, that is an attribute of a record, not recordset.21:52
KiallOh, humm..21:52
KiallWe probably want to extent storage to support that, but maybe file a Launchpad BP for it and do it in code on the mDNS side for the moment?21:53
Kiallextend*21:53
rjrjrsomething like criterion={"record.action": "!DELETE"} would work.21:54
rjrjris that what you are thinking?21:54
rjrjrabout how to extend storage...21:55
KiallYea, something along those lines21:59
KiallBut.. We can skip it for right now, file a BP and log a Bug/BP for later..21:59
KiallmDNS needs to iterate the list anyway for other reasons, so adding a if action = 'DELETE': continue is pretty harmless22:00
openstackgerritKiall Mac Innes proposed openstack/python-designateclient: Refactor KS session creation and support CA certs  https://review.openstack.org/13747922:01
openstackgerritKiall Mac Innes proposed openstack/python-designateclient: Refactor KS session creation and support CA certs  https://review.openstack.org/13747922:03
rjrjrkiall: okay.22:11
rjrjrkiall: just so i understand, right now, records are removed from the tables, but recordsets stick around.22:11
rjrjrwhen we introduce IXFR, we'll need records to stick around (at least for a while.)22:12
openstackgerritGraham Hayes proposed openstack/designate: Added designate-dashboard code to contrib/designate-dashboard  https://review.openstack.org/13748022:12
Kiallrjrjr: ^22:12
rjrjrcool.  dashboard code.  i'll let PK know.22:13
rjrjrso, back to what i was saying.  if i remove a record, even the last record in a recordset, the recordset is not removed.  does that sound correct?22:13
mugsieno... but it is known bug22:14
KiallI think recordsets sticking around was a bug, a bug I thought was fixed recently?22:14
mugsiei dont think we ever hads code to fix it22:14
mugsiebut I could be wrong22:14
rjrjri might have re-introduced it with what i'm working on.22:15
rjrjri moved the removal of records to the update_status method.22:15
rjrjr(from the table.)22:15
mugsieah, that could  have done it22:15
rjrjrbesides, we'll want records and recordsets to stick around when we work on the IXFR code.22:15
KiallHumm, is that for IXFR support?22:16
rjrjrno.22:16
mugsierjrjr: yeah, they will get the status column then22:16
mugsieand the "deleted" col22:16
rjrjri have notes to remove the 'deleted' column since it is redundant with status='DELETED'.  but, for now, it is not hurting anything.22:18
rjrjrold code - in the delete_record method we removed the record from the table.  new code - action is changed to 'DELETE' and status is 'PENDING'.  update_status actually removes the record if a threshold of servers agree it is deleted.22:18
rjrjr(i added a TODO about this to the code.  and have notes to fix this in the IXFR code.)22:18
rjrjrwith IXFR, we are not going to want to remove the record from the table (also have this in a TODO and will fix this in the IXFR code.)22:19
rjrjri have to update the IXFR spec, so i'll make sure i capture all this in that.22:20
rjrjrso, i thought i had things working well until i realized the records were not actually being removed from the zone.  that's what started me down this whole path...22:21
Kialldeleted column is actually quite important still :)22:24
KiallIt allows for the unique constraints to work correctly22:25
rjrjrgood to know.22:25
Kiallre "update_status actually removes the record if a threshold of servers agree it is deleted." I assume you mean updates status = COMPLETE and sets deleted = $id, or actually purges from the DB?22:25
rjrjrthe behavior is the same as before.  when the update_status is called and status is 'SUCCESS' the record that has an action of 'DELETE' and a status of 'PENDING', i call the Storage delete_record method.  i do the same for domains.22:28
rjrjrthe end result, i'm calling the same method that would have been called, just later in the process.22:29
rjrjrmugsie: thanks for adding the synchronize domain code by the way.  it was on my list of things to do, but now that it is already done.22:29
KiallK - I'm still wondering a little about the delete_record call being in the update_status method, I'm wondering how that might affect the V1 API22:30
rjrjrkiall: if you look at what i checked in, i have an if statement testing which backend is called in the delete_record method.22:31
rjrjrif we are using the pool_manager_proxy backend, we defer deleting the record until update_status.22:32
rjrjrif we use any other backend, we follow the same logic it always had.22:32
rjrjrit's easier to see this in code than for me to explain i think.22:33
rjrjrthat code is all there.22:33
KiallK - I'll test it out in the morning and have a better understanding :)22:34
rjrjrthe good news, the bulk of the code is complete.  i'm just debugging.22:35
rjrjri added the delay that was needed between notify_zone_changed and poll_for_serial_number.22:35
Kiallrjrjr: BTW - If your using the Vagrant VM to test our the Horizon bits etc, you'll need to bump the ram for the VM way up (say, 8GB) since all the openstack services need to be installed for Horizon not to explode22:35
rjrjri found this bug with the delete record not working (it doesn't remove the record from the zone in DNS.)22:36
Kiall:'(22:36
rjrjrand i also know about one more bug where the NS,SOA records do not have the correct status/action during the process and also have a serial number of 1.22:37
rjrjri am fixing all these now and will submit shortly.22:37
Kialltimsim: http://paste.openstack.org/show/85814/ (vinod / betsy - if your in the office with him, can you pass it on :))22:38
Kiallrjrjr: cool22:38
Kialltimsim: logs of conversation are http://eavesdrop.openstack.org/irclogs/%23openstack-dns/%23openstack-dns.2014-07-09.log22:38
rjrjrack concerning the VM ram.22:38
rjrjrone final question, did we actually use domain or record status before?22:40
rjrjrbefore this pool manager code?22:41
Kiallmugsie added them as an initial step towards a pre-precursor to pools.. So, No, I don't believe so.22:41
rjrjri see something where we test for PENDING in the API, but I didn't see much else for status.22:41
rjrjrcool.  i was worried about breaking something else, but i could find what else i could break.22:42
rjrjrwell, have a good night/day.  despite it being a 4 day weekend, i'll be working so we can get this pool manager code complete.  so if you review the code, i'll be checking all weekend.22:43
Kiall:)22:50
*** RaginBajin has quit IRC22:51
vinodKiall: timsim will be at work on Monday - i'll pass on the code to him then22:53
KiallPfft, taking the holidays off? Unacceptable ;)22:53
vinodwill let him know ;-)22:56
Kiallvinod: actually! Any chance of a quick review of https://review.openstack.org/#/c/137479/22:57
KiallIt's .. umm.. blocking some HP work and I need to cut a release with the CA Cert support ;)22:57
KiallIdeally before everyone goes away on holidays :P22:57
vinodlooking at it now23:03
KiallExcellent, Thanks :)23:03
openstackgerritKiall Mac Innes proposed openstack/python-designateclient: Refactor KS session creation and support CA certs  https://review.openstack.org/13747923:07
KiallTiny mistake of course ;)23:08
*** ryanpetrello has joined #openstack-dns23:12
openstackgerritGraham Hayes proposed openstack/designate: Added designate-dashboard code to contrib/designate-dashboard  https://review.openstack.org/13748023:18
mugsiedamn pep823:18
*** ryanpetrello has quit IRC23:19
*** jmcbride1 has joined #openstack-dns23:21
*** jmcbride has quit IRC23:24
vinodKiall: you have a +2 from me assuming the gate checks pass23:26
vinodhmm. close to midnight for you guys23:27
KiallFU**.23:27
KiallMissed the train home -_-23:27
mugsieshit23:27
mugsiedamn you bad public transport in ireland23:28
vinodso sorry to hear that - does not sound good at all23:30
vinodmost of us won't be around till Monday. if you need anything to be looked at before Monday - send me an email.23:33
KiallYep, enjoy the time off :P23:34
*** vinod has quit IRC23:34
*** GonZoPT has joined #openstack-dns23:47
*** GonZo2K has quit IRC23:49
*** rediskin has left #openstack-dns23:51

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