Tuesday, 2017-08-08

*** tongl has quit IRC00:16
*** livelace has quit IRC01:28
*** livelace has joined #openstack-neutron-ovn01:29
*** mestery has quit IRC02:17
*** afranc has quit IRC02:17
*** mestery has joined #openstack-neutron-ovn02:22
*** afranc has joined #openstack-neutron-ovn02:22
*** saphi has joined #openstack-neutron-ovn02:24
*** saphi has quit IRC02:38
*** yamamoto_ has joined #openstack-neutron-ovn02:46
*** yamamoto has quit IRC02:46
*** saphi has joined #openstack-neutron-ovn02:51
*** anilvenkata has joined #openstack-neutron-ovn03:56
*** janki has joined #openstack-neutron-ovn05:03
*** saphi_ has joined #openstack-neutron-ovn05:24
*** saphi has quit IRC05:25
*** lrichard has quit IRC05:47
*** lrichard has joined #openstack-neutron-ovn05:53
*** saphi__ has joined #openstack-neutron-ovn06:01
*** saphi_ has quit IRC06:02
*** livelace has quit IRC06:22
*** livelace has joined #openstack-neutron-ovn06:28
*** mmirecki has joined #openstack-neutron-ovn06:45
*** zefferno has joined #openstack-neutron-ovn06:57
*** pcaruana has joined #openstack-neutron-ovn07:00
*** gtrxcb has quit IRC07:11
*** yamamoto_ has quit IRC07:44
*** yamamoto has joined #openstack-neutron-ovn07:50
*** mwarad has joined #openstack-neutron-ovn08:15
*** _mwarad_ has joined #openstack-neutron-ovn08:15
*** _mwarad_ has quit IRC08:15
*** Guest14 has joined #openstack-neutron-ovn08:21
*** lucas-afk is now known as lucasagomes08:26
openstackgerritJakub Libosvar proposed openstack/ovsdbapp master: Add RowView versions of db_find and db_list  https://review.openstack.org/49019908:32
openstackgerritJakub Libosvar proposed openstack/ovsdbapp master: Add RowView versions of db_find and db_list  https://review.openstack.org/49019908:33
*** yamamoto has quit IRC08:49
*** mwarad has quit IRC08:59
*** yamamoto has joined #openstack-neutron-ovn09:15
*** yamamoto has quit IRC09:22
*** pmannidi has joined #openstack-neutron-ovn09:48
openstackgerritMerged openstack/networking-ovn master: Prepare for using ovsdbapp > 0.4.0  https://review.openstack.org/48744209:50
*** yamamoto has joined #openstack-neutron-ovn09:54
*** saphi__ has quit IRC10:01
*** yamamoto has quit IRC10:07
*** yamamoto has joined #openstack-neutron-ovn10:12
*** yamamoto has quit IRC10:17
*** yamamoto has joined #openstack-neutron-ovn10:18
*** yamamoto has quit IRC10:20
*** yamamoto has joined #openstack-neutron-ovn10:20
openstackgerritMiguel Angel Ajo proposed openstack/networking-ovn master: refarch: Update documentation and diagrams  https://review.openstack.org/49081210:54
Guest14dalvarez can you have a look on metadata related changes here ^ ?10:54
*** Guest14 is now known as ajo10:54
ajoyikes10:54
dalvarezajo sure10:55
ajothanks :D10:56
*** yamamoto has quit IRC10:58
*** yamamoto has joined #openstack-neutron-ovn10:59
*** yamamoto has quit IRC11:05
*** yamamoto has joined #openstack-neutron-ovn11:07
*** yamamoto has quit IRC11:08
*** yamamoto has joined #openstack-neutron-ovn11:10
*** yamamoto has quit IRC11:13
*** yamamoto has joined #openstack-neutron-ovn11:15
*** yamamoto has quit IRC11:16
*** yamamoto has joined #openstack-neutron-ovn11:16
openstackgerritLucas Alvares Gomes proposed openstack/networking-ovn master: Idea proposal: Neutron/OVN database consistency problem  https://review.openstack.org/49083411:38
openstackgerritLucas Alvares Gomes proposed openstack/networking-ovn master: Tox docs: Force the use of python2.7  https://review.openstack.org/49176511:43
*** lucasagomes is now known as lucas-hungry11:46
*** yamamoto has quit IRC12:18
*** yamamoto has joined #openstack-neutron-ovn12:20
*** yamamoto has quit IRC12:20
openstackgerritMerged openstack/networking-ovn master: call provisioning_complete conditionally  https://review.openstack.org/49051112:33
*** yamamoto has joined #openstack-neutron-ovn12:40
*** lucas-hungry is now known as lucasagomes12:45
ajorussellb , anilvenkata about the check db version / vs model introspection in networking-ovn12:51
ajoI was proposing more the introspection side, while, it could be marginally slower / or a bit more complex to program12:51
ajobecause it's a bit more independent of backports12:51
ajowhen maintaining neutron upstream I cried many times because it wasn't easy to backport a fix that was db dependent12:52
ajomaking sure we don't use the version for checks would avoid us eventually ending in the same situation12:52
anilvenkataajo, how would using version will affect backporting issues?12:53
ajomay be it's over engineering at this point, I don't know, but I remember hours of frustration trying to find a solution for the db backports, that... while they ended being possible, they were technically very complicated to handle12:53
ajoanilvenkata , you backport an specific patch from 2.9 to 2.8, for example, that exposes a new column/table via the db12:54
ajowhile you cannot set the 2.9 version of the DB (because that would misslead later upgrades)12:54
ajothe colum/table would still be there12:54
ajoimagine that you had 5.8.1 in 2.812:54
ajoand in 2.9 the patch you try to backport upgrades db from 5.9.3 to 5.9.412:55
ajoyou could backport the patch and...12:55
ajoupdate the db in 2.8 to 5.8.1.212:55
ajoor someyhing like that12:55
openstackgerritMiguel Angel Ajo proposed openstack/networking-ovn master: refarch: Update documentation and diagrams  https://review.openstack.org/49081212:57
anilvenkataajo, in this case our version check in networking-ovn will be, before backporting >=5.9.4 , after backporting  >=5.8.1.212:57
anilvenkataajo, can't we solve backporting issule like that ^ ?12:58
numansajo anilvenkata how can you backport something if the corresponding DB table/column is not present in the older version ?12:58
ajonumans backporting the db changes too13:00
anilvenkatanumans, ajo I think in backporting patch, he is also backporting DB changes13:00
anilvenkata:)13:00
ajoanilvenkata you're right, we could change the check for the DB version13:00
ajoto the new 5.8.1.213:00
ajowell13:01
ajothat'd make sense13:01
ajoanilvenkata++13:01
ajook, so let it be db version checks :)13:01
anilvenkataajo, no still that is a problem13:01
ajowhy13:01
anilvenkataajo, what if the user is not upgrading new networking-ovn13:02
ajothen they wouldn't be using the new feature/thing13:02
*** mmichelson has joined #openstack-neutron-ovn13:02
ajonorthd interface would have to be backwards compatible (like we did for the options:redirect_chassis & Gateway_Chassis)13:03
ajoovn-northd will take any of them13:03
anilvenkataajo, but if the new feature deletes columns?13:03
ajowith priority for the new13:03
ajoanilvenkata that'd be an issue in either case13:03
ajoif you check the models or if you check the version13:03
ajoas soon as a column you arent' checking goes away13:03
ajoI guess deleting things from the DB can't be done lightly13:04
ajoit needs a deprecation window13:04
anilvenkatarussellb, any thoughts ^ on if we can go with version number or other approach?13:05
anilvenkataajo, russellb if we can conclude, I will update my patch :)13:05
numansajo, got it.13:05
numansanilvenkata, ajo, although I feel its highly unlikely that a feature/db changes would be backported  to a released branch13:05
ajoI think, based on the discusion, version check may be ok13:05
ajonumans yeah, it's always the case, until you find it13:06
ajo:D13:06
ajobut, it could be fixed with versioning to13:06
ajotoo13:06
ajoyou just change the check for the "new backported version" is not a big deal :)13:06
anilvenkataajo, thanks Miguel13:07
anilvenkataajo, russellb numans shall I go ahead with version checks?13:07
numansanilvenkata, version checks seems simpler in terms of code. so i am fine with it :)13:10
anilvenkatanumans, thanks numans13:11
russellbi'm really ok with either ... i guess it's more my ovn maintainer hat on, thinking "why do we bother with the version if it doesn't help solve a problem"13:13
anilvenkatarussellb, thanks Russell13:19
russellbbut i also get how backports make version checking messy, so checking for the feature is the most explicit13:20
anilvenkatarussellb, checking for feature means checking if columns and tables present or not?13:35
anilvenkataajo, ^13:35
anilvenkatanumans, ^13:37
russellbyes13:40
*** mlavalle has joined #openstack-neutron-ovn13:43
anilvenkatarussellb, thanks Russell13:45
*** anilvenkata is now known as anilvenkata|AFK13:45
numansanilvenkata|AFK, even i was thinking the same. checking for the feature seems to be better13:49
*** anilvenkata|AFK is now known as anilvenkata13:58
anilvenkatanumans, thanks Numa13:59
*** pmannidi has quit IRC14:14
*** mmirecki has quit IRC15:05
*** mmirecki has joined #openstack-neutron-ovn15:25
*** livelace2 has joined #openstack-neutron-ovn15:48
*** janki has quit IRC15:48
*** mmirecki has quit IRC15:51
*** yamamoto has quit IRC15:54
*** zefferno has quit IRC15:57
*** ajo has quit IRC15:57
*** yamamoto has joined #openstack-neutron-ovn15:58
*** yamamoto has quit IRC16:03
openstackgerritLucas Alvares Gomes proposed openstack/networking-ovn master: Idea proposal: Neutron/OVN database consistency problem  https://review.openstack.org/49083416:23
*** lucasagomes is now known as lucas-afk16:24
openstackgerritTerry Wilson proposed openstack/ovsdbapp master: Add RowView versions of db_find and db_list  https://review.openstack.org/49019916:25
*** pcaruana has quit IRC16:27
*** saphi has joined #openstack-neutron-ovn16:32
openstackgerritvenkata anil proposed openstack/networking-ovn master: Support for L3 gateway HA  https://review.openstack.org/48697116:48
*** anilvenkata has quit IRC16:53
*** saphi has quit IRC16:56
*** livelace2 has quit IRC17:55
*** mlavalle has quit IRC19:00
*** yamamoto has joined #openstack-neutron-ovn21:09
*** yamamoto has quit IRC21:21
*** yamamoto has joined #openstack-neutron-ovn21:59
*** mmichelson has quit IRC22:24

Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!