Friday, 2014-02-21

*** sn6i23a has quit IRC00:16
*** jorgem has quit IRC00:17
*** CaptTofu has joined #openstack-dns00:23
*** matsuhashi has joined #openstack-dns00:31
*** matsuhas_ has joined #openstack-dns00:48
*** matsuhashi has quit IRC00:55
*** openstackgerrit has quit IRC01:18
*** openstackgerrit has joined #openstack-dns01:18
*** ChanServ sets mode: +v openstackgerrit01:18
*** nosnos has joined #openstack-dns01:34
*** shakayumi has quit IRC01:53
*** rossk has quit IRC01:59
*** jmcbride has joined #openstack-dns02:09
*** elemecca1 has quit IRC02:12
*** jmcbride has quit IRC02:21
*** matsuhas_ has quit IRC02:34
*** matsuhas_ has joined #openstack-dns02:36
*** elemecca has joined #openstack-dns02:43
*** elemecca has quit IRC02:48
*** vinod has joined #openstack-dns02:54
*** matsuhas_ has quit IRC03:03
*** matsuhas_ has joined #openstack-dns03:06
*** vinod has quit IRC03:25
*** matsuhas_ has quit IRC03:29
*** richm has quit IRC03:43
*** CaptTofu has quit IRC04:02
*** furlongm has quit IRC04:04
*** msisk has joined #openstack-dns04:23
*** matsuhashi has joined #openstack-dns04:25
*** msisk has quit IRC04:38
*** jorgem has joined #openstack-dns04:52
*** jorgem has quit IRC04:53
*** jorgem has joined #openstack-dns04:53
*** matsuhashi has quit IRC05:23
*** jmcbride has joined #openstack-dns05:29
*** matsuhashi has joined #openstack-dns05:33
*** jmcbride has quit IRC05:34
*** CaptTofu has joined #openstack-dns06:03
*** CaptTofu has quit IRC06:08
*** jorgem1 has joined #openstack-dns06:47
*** jorgem has quit IRC06:48
*** jorgem1 has quit IRC06:48
*** jorgem has joined #openstack-dns06:48
*** jorgem1 has joined #openstack-dns07:03
*** jorgem has quit IRC07:05
*** jorgem has joined #openstack-dns07:09
*** jorgem1 has quit IRC07:13
*** CaptTofu has joined #openstack-dns08:04
*** CaptTofu has quit IRC08:09
*** jorgem has quit IRC08:23
*** jorgem has joined #openstack-dns08:23
*** CaptTofu has joined #openstack-dns10:05
*** CaptTofu has quit IRC10:09
*** CaptTofu has joined #openstack-dns10:57
*** nosnos has quit IRC10:58
*** matsuhashi has quit IRC11:02
*** jmcbride has joined #openstack-dns11:29
*** jmcbride has quit IRC11:36
*** mwagner_lap has joined #openstack-dns13:11
*** CaptTofu has quit IRC13:19
*** mwagner_lap has quit IRC13:38
*** jmcbride has joined #openstack-dns13:44
*** nkinder has quit IRC14:13
*** eankutse has joined #openstack-dns14:24
*** rektide_ is now known as rektide14:41
*** CaptTofu has joined #openstack-dns14:50
*** vinod has joined #openstack-dns14:59
vinodkiall: are you around?15:05
*** richm has joined #openstack-dns15:06
*** nkinder has joined #openstack-dns15:06
kiallvinod: Heya15:36
vinodWanted to talk about the comments that you had regarding specific exceptions - https://review.openstack.org/#/c/7514315:38
vinodTo throw specific exceptions, I will need to duplicate the sort_dir check in storage15:38
kiallvinod: so, my thinking was, if a user gets a "value error", that's really confusing15:39
vinod            # Limits are checked at the API layer.  If however central or15:39
vinod            # storage is called directly, invalid limits show up as15:39
vinod            # ValueError15:39
vinodAt the API layer I can throw InvalidLimit15:39
kiallhttps://review.openstack.org/#/c/75143/2/designate/api/v2/controllers/rest.py15:39
kiallL63 - could be switched to raise exceptions.InvalidLimit()15:40
vinodYes I will be doing that15:40
kiallL65 to exceptions.InvalidSort()15:40
kiallWould that not cover it?15:40
vinodL65 should also be exceptions.InvalidLimit15:41
vinodHere we are checking only for invalid limit15:41
vinodsort_dir, marker and sort_key are checked here - https://review.openstack.org/#/c/75143/2/designate/storage/impl_sqlalchemy/__init__.py15:41
vinodIdeally I would like all the checks to be in storage at https://review.openstack.org/#/c/75143/2/designate/storage/impl_sqlalchemy/__init__.py15:42
vinodBut all invalid limits are not caught there - and especially the rpc limit cannot be15:42
vinodso i moved just the limit checks to https://review.openstack.org/#/c/75143/2/designate/api/v2/controllers/rest.py15:42
*** msisk has joined #openstack-dns15:42
vinodhttp://pastebin.com/E3DZHXhD15:44
kiallAh .. Humm .. Well, sort_dir and sort_key could be checked in the API15:47
kiallsort_dir is easy enough .. it's always asc or desc15:47
kiallsort_key could be checked against a per-controller list of valid sort keys15:48
kiallwhich gives a added bonus of allowing us to decide which keys can be used to sort against15:48
kiall(e.g. we probably don't want people sorting records by there value.. that's a sql blob after all)15:48
vinodhaving the checks in controller, implies, if you call central then the same checks are not done15:51
*** artom has joined #openstack-dns15:51
kiallWell - Most of our validation is already done like that (I think theres a BP or Bug re moving it around a bit)15:51
kiallI'm generally a fan of validating at the edges, and not letting crummy values get past the outermost layer.15:53
kiallThe downside to that is.. When the validation isn't abstracted, you're stuck duplicating in in places.. Which sucks.15:53
kiallbrb15:53
artomkiall, could validation moved into its own component/module?15:57
kiallSo - I've been thinking about the pluggable record type stuff, which kinda merges into the structured rdata format stuff16:02
kiallI had a review up (it's expired..) that added classes for each of the record types16:02
kiallIf we extended that to all our core objects.. Zones, RecordSets etc16:02
kiallthen validations for everything have a neat home that can be used by all the layers16:02
vinodi will look at your review again - but from an initial glance it is a good idea16:11
vinodfor now i will do the validations that i can in the api layer16:11
*** jmcbride has quit IRC16:12
*** msisk_ has joined #openstack-dns16:14
*** msisk has quit IRC16:16
*** CaptTofu has quit IRC16:16
*** jmcbride has joined #openstack-dns16:17
kiallvinod: yea, I think it's the right way to fix it, we only really thought of making the switch for everything a day or two ago.. So haven't had time to actually see how much effort it would be16:19
kiallI suspect not much .. Since we can make our Objects behave just like the dicts and lists our Storage layer returns today... Hopefully meaning virtually no changes outside the storage layer for the initial switch16:20
*** vinod has quit IRC16:26
*** jmcbride has quit IRC16:27
*** CaptTofu has joined #openstack-dns16:30
*** vinod has joined #openstack-dns16:32
*** jmcbride has joined #openstack-dns16:34
ekarlsokiall: yo16:35
ekarlsoyou guys around ?16:36
kiallYep16:36
ekarlsowhat's the paging fuzz about ?16:37
kiallYou missed a few things ;)16:37
kiall(And we missed them in testing)16:38
kiallRead the commit, you'll see16:38
ekarlsokk16:39
ekarlsoa q though16:40
ekarlsosince we have a storage layer16:40
ekarlsowhy don't we have a model layer since we're supposed to have "pluggable" storage?16:40
kialllol .. scroll up about 10 lines16:41
ekarlsooh ok16:41
ekarlsoyou working on thaet ?16:41
kiallNobody is working on it .. Because we haven't decided A) if it's a good idea (Okay, it probably is) and B) if the change should be rushed into icehouse (It probably shouldn't be)16:44
vinod#agree on doing this after icehouse16:46
ekarlsoyeah, do we semi-know what approach we wanna have ?16:47
ekarlsoas in just models or the nova objects thing16:47
kiallekarlso: please don't go starting it ;)16:47
ekarlsokiall: no, I won't, just wondering to make a "initial bp" for juno16:47
ekarlsothen at least it's noted down somewher16:48
kiallNo approach has been thought through .. other than 15:52 <kiall> I had a review up (it's expired..) that added classes for each of the record types16:48
ekarlsook : )16:48
kiallBut - It didn't really aim to be this, it just triggered the idea16:48
ekarlsoOk if I add a BP though in draft so we have it noted down somewhere?16:49
kiallNo, not right now16:49
ekarlsook ! :)16:50
kiallI'll write up a BP once the idea's actually been thought through a little more than "Gee, this might just solve some problems for us" ;)16:50
ekarlso:)16:50
kiallYou get anywhere with bug16:51
kiall#1282672 today?16:51
ekarlsoyeah16:51
ekarlsogonna push in an our : )16:51
ekarlsojust adding tests also tio it : p16:51
*** jmcbride has quit IRC16:52
*** jmcbride has joined #openstack-dns16:52
vinodekarlso: is 1282672 different from 1282682?16:52
vinodIs this for checking the UUIDs in the actual URL as opposed to marker?16:53
ekarlsovinod: yes16:53
*** betsy has joined #openstack-dns16:55
*** nkinder has quit IRC17:04
*** jmcbride has quit IRC17:18
*** nkinder has joined #openstack-dns17:20
*** jorgem has joined #openstack-dns17:24
*** elemecca has joined #openstack-dns17:36
*** jmcbride has joined #openstack-dns18:04
vinodkiall: In the api tests, is there a way to assert that I need to get an InvalidLimit exception (from the api and not central), rather than just a 400?18:14
vinodwith testtools.ExpectedException() fails18:15
vinodAssertionError: InvalidLimit not raised.18:15
kiallYou can check the type filed in the response == 'invalid_limit'18:24
kiallBut, the API catches the exceptions and converts them.. so assertRaises and ExpectedException won't work18:25
kiallfield*18:25
*** tsimmons has joined #openstack-dns18:25
tsimmonskiall: How would you go about writing out the bind configs in the devstack script?18:27
*** msisk_ has quit IRC18:27
kialltsimmons: heya18:28
kiallIt depends really, if the files are going to be 99% the same on Ubuntu 12, Ubuntu 14, Fedora 18, Fedora 19 etc etc, then a bash HEREDOC with some variables for path differences etc should work18:30
kiallIf there really different, then a if is_ubuntu; then render HEREDOC; else; EXPLODE; fi would be better, as it will at least shout at people that their OS needs to be implemented18:31
tsimmonsRight, I've never tried to install bind on anything other than Ubuntu/Debian. Not sure how it differs.18:31
kiallAh .. Then go with the "blow up if it's not ubuntu" pattern, and the current stuff (or rendering the entire config) should do the trick ;)18:33
tsimmonsAlright, will do :) i don't think it's much different, I could look into it, but I'll fix that up. Thank you :)18:33
kiallGotta run.. Cyas later18:34
tsimmonsSee ya, thanks again.18:34
*** jmcbride has quit IRC18:34
*** jmcbride has joined #openstack-dns18:35
openstackgerritVinod Mangalpally proposed a change to stackforge/designate: Handle invalid pagination parameters  https://review.openstack.org/7514318:48
*** tsimmons has quit IRC19:00
*** jorgem1 has joined #openstack-dns19:05
*** jorgem has quit IRC19:08
*** msisk has joined #openstack-dns19:22
*** msisk has quit IRC19:23
*** msisk has joined #openstack-dns19:33
*** msisk has quit IRC19:34
*** jmcbride has quit IRC19:42
*** sn6i23a has joined #openstack-dns19:47
*** vinod has quit IRC20:12
*** jmcbride has joined #openstack-dns20:13
*** msisk has joined #openstack-dns20:13
*** msisk has quit IRC20:14
*** msisk has joined #openstack-dns20:14
*** tsimmons has joined #openstack-dns20:20
*** jorgem1 has quit IRC20:28
openstackgerritTim Simmons proposed a change to stackforge/designate: Add Bind9 Support to the DevStack Plugin  https://review.openstack.org/7522220:39
*** jorgem has joined #openstack-dns20:40
*** vinod has joined #openstack-dns20:42
openstackgerritTim Simmons proposed a change to stackforge/designate: Add Bind9 Support to the DevStack Plugin  https://review.openstack.org/7522220:48
*** ekarlso has quit IRC20:50
*** ekarlso has joined #openstack-dns20:50
*** eankutse has quit IRC21:06
*** tsimmons has quit IRC21:11
*** tsimmons has joined #openstack-dns21:25
*** jmcbride1 has joined #openstack-dns21:27
*** jmcbride has quit IRC21:27
*** tsimmons has left #openstack-dns21:30
*** vinod1 has joined #openstack-dns21:36
*** vinod has quit IRC21:36
*** vinod has joined #openstack-dns21:38
*** vinod1 has quit IRC21:41
*** CaptTofu has quit IRC22:10
*** msisk has quit IRC22:12
*** CaptTofu has joined #openstack-dns22:12
*** msisk has joined #openstack-dns22:13
*** jorgem has quit IRC22:39
*** jorgem has joined #openstack-dns22:41
*** vinod has quit IRC22:42
*** betsy has quit IRC22:44
*** jmcbride1 has quit IRC22:47
*** CaptTofu has quit IRC22:54
*** jmcbride has joined #openstack-dns23:03
*** jorgem1 has joined #openstack-dns23:07
*** jorgem1 has quit IRC23:08
*** jorgem has quit IRC23:09
*** artom has quit IRC23:11
*** richm has quit IRC23:25
*** nkinder has quit IRC23:38
*** richm has joined #openstack-dns23:39
*** msisk has quit IRC23:59

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