Friday, 2014-03-14

openstackgerritAaron Rosen proposed a change to openstack/neutron: NSX: remove last of unneed quantum references  https://review.openstack.org/8038800:00
*** leseb has joined #openstack-neutron00:01
*** hemanthravi has joined #openstack-neutron00:01
*** pasquier-s has quit IRC00:02
*** pasquier-s has joined #openstack-neutron00:02
*** leseb has quit IRC00:05
*** pcm_ has quit IRC00:08
*** SumitNaiksatam has quit IRC00:11
*** ramishra has quit IRC00:15
*** emagana has quit IRC00:15
*** gdubreui has quit IRC00:15
*** jecarey has quit IRC00:18
*** jecarey_ has joined #openstack-neutron00:18
openstackgerritAaron Rosen proposed a change to openstack/neutron: Skip radware failing test for now  https://review.openstack.org/8042600:18
openstackgerritAaron Rosen proposed a change to openstack/neutron: NSX: Fix router-interface-delete returns 404 when router not in nsx  https://review.openstack.org/8017700:18
openstackgerritAaron Rosen proposed a change to openstack/neutron: Skip radware failing test for now  https://review.openstack.org/8042600:18
*** ramishra has joined #openstack-neutron00:19
*** ramishra has quit IRC00:19
*** ramishra has joined #openstack-neutron00:19
openstackgerritAaron Rosen proposed a change to openstack/neutron: Notify nova when ports are ready  https://review.openstack.org/7525300:20
openstackgerritAaron Rosen proposed a change to openstack/neutron: Send network-changed notifications to nova  https://review.openstack.org/7804100:20
*** _cjones__ has quit IRC00:22
*** hemanthravi has quit IRC00:24
*** gdubreui has joined #openstack-neutron00:27
*** networkstatic has quit IRC00:27
*** hemanthravi has joined #openstack-neutron00:27
*** networkstatic has joined #openstack-neutron00:27
*** matsuhashi has joined #openstack-neutron00:31
*** ramishra has quit IRC00:31
*** thedodd has joined #openstack-neutron00:36
*** suresh12_ has quit IRC00:38
*** suresh12 has joined #openstack-neutron00:40
*** tomoe_ has joined #openstack-neutron00:45
*** thuc has quit IRC00:45
*** thuc has joined #openstack-neutron00:45
*** openstack has joined #openstack-neutron00:53
*** dims_ has joined #openstack-neutron00:53
markmcclainbecause it is a new feature which was not merged by Icehouse-300:53
*** joshh has quit IRC00:55
*** nati_uen_ has quit IRC00:55
*** tomoe_ has quit IRC00:55
*** openstack has joined #openstack-neutron00:57
*** manishg has quit IRC00:57
*** openstack has quit IRC00:57
*** openstack has joined #openstack-neutron00:59
*** samuelbercovici has quit IRC00:59
*** yfujioka has quit IRC01:01
*** samuelbercovici1 has quit IRC01:01
*** harmw has quit IRC01:01
*** beagles has quit IRC01:01
*** jgrimm has quit IRC01:02
*** samuelbercovici has joined #openstack-neutron01:02
*** yfujioka_ has joined #openstack-neutron01:02
*** rook has quit IRC01:02
*** ywu has quit IRC01:02
*** harmw has joined #openstack-neutron01:02
*** ywu has joined #openstack-neutron01:02
*** yfujioka has joined #openstack-neutron01:02
*** yfujioka_ has quit IRC01:03
*** SpamapS has quit IRC01:04
*** ywu has quit IRC01:07
*** b3nt_pin has joined #openstack-neutron01:07
*** pasquier-s has quit IRC01:07
*** pasquier-s has joined #openstack-neutron01:08
*** gdubreui has joined #openstack-neutron01:08
*** dguitarbite has quit IRC01:08
*** matsuhashi has quit IRC01:08
*** SpamapS_ has joined #openstack-neutron01:08
*** thedodd has quit IRC01:08
*** ywu has joined #openstack-neutron01:08
*** baoli has joined #openstack-neutron01:08
*** baoli has quit IRC01:10
*** baoli has joined #openstack-neutron01:10
*** julim has quit IRC01:12
*** matsuhashi has joined #openstack-neutron01:12
*** openstack has joined #openstack-neutron01:15
*** openstack has joined #openstack-neutron01:19
*** rm_work has joined #openstack-neutron01:19
*** kedarkul has joined #openstack-neutron01:19
*** rm_work has quit IRC01:19
*** rm_work has joined #openstack-neutron01:19
*** harmw has quit IRC01:24
*** openstack has quit IRC01:24
*** openstack has joined #openstack-neutron01:25
*** joshh has quit IRC01:26
*** pasquier-s has quit IRC01:26
*** openstack has joined #openstack-neutron15:09
*** vjay has joined #openstack-neutron15:09
*** yamahata has quit IRC15:14
*** yamahata has joined #openstack-neutron15:14
*** devvesa has quit IRC15:14
*** dvorkinista has quit IRC15:15
*** dvorkinista has joined #openstack-neutron15:15
*** networkstatic has quit IRC15:17
*** networkstatic has joined #openstack-neutron15:17
*** rohit404 has joined #openstack-neutron15:19
*** sacharya has quit IRC15:19
*** HenryG has quit IRC15:20
*** emagana has joined #openstack-neutron15:22
*** claudiub___ has joined #openstack-neutron15:24
*** ijw has joined #openstack-neutron15:26
openstackgerritBob Melander proposed a change to openstack/neutron: Implements Neutron routing API using Cisco appliances  https://review.openstack.org/7419415:28
*** dvorkinista has quit IRC15:30
*** HenryG has joined #openstack-neutron15:35
matrohonmarkmcclain : hi15:36
markmcclainmatrohon: hi15:37
matrohonmarkmcclain : I just wanted to know exactly what you meant when you said there is no test outside unit test during this review : https://review.openstack.org/#/c/49227/15:38
matrohonmarkmcclain : do you mean integration test with tempest?15:39
markmcclainmatrohon: the l2-pop driver is not tested via a 3rd party system or in the gate15:39
*** marun has quit IRC15:39
markmcclainso we have no data whether this works outside of developer testing15:40
markmcclainwhen Juno opens15:40
markmcclainI was thinking that l2pop should be one of the configurations we test in the check queue for neutron15:40
matrohonmarkmcclain : yes and we are working on it. But infra reject our evolution of devstack to have multi-node with lxc :15:41
matrohonhttps://review.openstack.org/#/c/76861/15:41
markmcclainmatrohon: can we enable it in a single vm devstack test15:41
markmcclainand get some functional coverage?15:42
matrohonmarkmcclain : yes, that was our goal15:42
openstackgerritAaron Rosen proposed a change to openstack/neutron: NSX: remove last of unneed quantum references  https://review.openstack.org/8038815:42
markmcclainI'd feel a lot better merging having a test that runs this in a single vm15:43
markmcclainI know the multi-node is best, but that will take time15:43
matrohonmarkmcclain : agree, and infra team is not clear about the good path to implement multi-node (node-pool evolution, tripleO, Heat..)15:44
*** thuc has joined #openstack-neutron15:44
*** devvesa has joined #openstack-neutron15:45
matrohonthe LXC config works in our lab, that the reason why we proposed it to infra guy15:45
markmcclainyeah15:46
*** otherwiseguy has quit IRC15:46
matrohonbut to them, and I understand there point, there are non-starter issue, du to iscsi module and ovs modules which are incompatible with a lxc config15:47
*** drumkilla is now known as rustlebee15:47
*** thuc_ has quit IRC15:48
*** thuc has quit IRC15:48
matrohonto me it would wonderful to have just a check job for neutron which launch a devstack with lxc computes, inside the devstack VM15:49
matrohonthat the reason why we proposed the devstack patch15:49
*** markmcclain has quit IRC15:50
*** arosen-home has joined #openstack-neutron15:50
*** sacharya has joined #openstack-neutron15:50
*** dave_tucker_zzz is now known as dave_tucker15:52
*** alagalah has joined #openstack-neutron15:53
*** sfox has joined #openstack-neutron15:57
openstackgerritClaudiu Belu proposed a change to openstack/neutron: Fixes the Hyper-V agent individual ports metrics  https://review.openstack.org/7879515:58
*** _cjones_ has joined #openstack-neutron16:03
*** jorgem has joined #openstack-neutron16:09
*** chandan_kumar has quit IRC16:14
*** rkukura has left #openstack-neutron16:16
*** pasquier-s_ has quit IRC16:17
*** salv-mobile has quit IRC16:17
*** thuc has joined #openstack-neutron16:18
*** thuc has quit IRC16:19
*** thuc has joined #openstack-neutron16:19
*** thuc has quit IRC16:20
*** jlibosva has quit IRC16:20
*** dansmith is now known as damnsmith16:20
*** thuc has joined #openstack-neutron16:21
matrohonmarkmcclain : l2-pop can't work with single VM, since we need tunneling16:22
*** jecarey has quit IRC16:22
*** jgallard has quit IRC16:23
*** nati_ueno has quit IRC16:24
*** vkozhukalov_ has joined #openstack-neutron16:26
*** jistr has quit IRC16:31
*** arosen-home has quit IRC16:32
*** dvorkinista has joined #openstack-neutron16:33
*** bjornar has joined #openstack-neutron16:35
*** tomoe_ has quit IRC16:37
*** openstackstatus has joined #openstack-neutron16:38
*** rkukura has joined #openstack-neutron16:44
*** dvorkinista has quit IRC16:48
*** dvorkinista has joined #openstack-neutron16:48
*** jgrimm has quit IRC16:50
openstackgerritA change was merged to openstack/neutron: Adds state reporting to SDN-VE agent  https://review.openstack.org/7883916:50
*** dims has joined #openstack-neutron16:54
*** matrohon has quit IRC16:55
*** SridharG has quit IRC16:56
*** petertoft has quit IRC16:57
*** dims has quit IRC17:00
*** dvorkinista has quit IRC17:00
*** markmcclain has joined #openstack-neutron17:00
*** claudiub___ has quit IRC17:00
*** markmcclain has quit IRC17:00
*** dvorkinista has joined #openstack-neutron17:00
*** markmcclain has joined #openstack-neutron17:01
*** nati_ueno has joined #openstack-neutron17:01
openstackgerritRossella Sblendido proposed a change to openstack/neutron: Introduce bulk calls for get device details  https://review.openstack.org/6689917:02
*** harlowja_away is now known as harlowja17:02
*** carl_baldwin_ has joined #openstack-neutron17:03
*** dvorkinista has quit IRC17:03
*** dvorkinista has joined #openstack-neutron17:03
*** suresh12 has joined #openstack-neutron17:04
*** carl_baldwin has quit IRC17:04
*** carl_baldwin_ is now known as carl_baldwin17:04
*** sphoorti has quit IRC17:05
*** jorgem has quit IRC17:06
*** jorgem1 has joined #openstack-neutron17:06
*** Guest58201 has joined #openstack-neutron17:07
Guest58201salut17:08
*** carl_baldwin_ has joined #openstack-neutron17:08
Guest58201j'ai un problème avec neutron17:08
Guest58201voilà le pastebin avec ma conf et quelques tests17:09
*** devvesa has quit IRC17:09
*** carl_baldwin has quit IRC17:09
*** carl_baldwin_ is now known as carl_baldwin17:09
Guest58201http://pastebin.com/tpRcNfX517:09
Guest58201et la topologie en place http://lut.im/htIQBnlE/loBTQJjh17:09
*** sfox has quit IRC17:10
*** claudiub___ has joined #openstack-neutron17:11
*** sfox has joined #openstack-neutron17:14
*** skraynev is now known as skraynev_afk17:14
*** SumitNaiksatam has joined #openstack-neutron17:16
*** alagalah has quit IRC17:16
*** otherwiseguy has joined #openstack-neutron17:18
*** afazekas has quit IRC17:18
*** alagalah has joined #openstack-neutron17:20
*** morganfainberg_Z is now known as morganfainberg17:21
markmcclainGuest58201: br-ex is not connected to anything17:22
Guest58201markmcclain: i have to connect br-ex ?17:27
Guest58201markmcclain: fr ?17:27
Guest58201:x17:27
*** jp_at_hp1 has joined #openstack-neutron17:28
*** jp_at_hp has quit IRC17:28
markmcclainGuest58201: I can read basic french :)17:28
openstackgerritClaudiu Belu proposed a change to openstack/neutron: Fixes the Hyper-V agent individual ports metrics  https://review.openstack.org/7879517:31
*** sfox1 has joined #openstack-neutron17:32
*** sfox has quit IRC17:33
markmcclainGuest58201: yes have to add an interface to the br-ex switch otherwise traffic won't flow17:34
*** jecarey has joined #openstack-neutron17:34
Guest58201which interface ?17:36
Guest58201like i want ?!17:36
Guest58201my data/vm network is 192.168.10.0/24 on eth217:36
*** ygbo has quit IRC17:38
*** safchain has quit IRC17:40
*** dims has joined #openstack-neutron17:40
*** leseb has quit IRC17:42
*** crc32 has joined #openstack-neutron17:44
Guest58201markmcclain: on eth2 its ok ?17:46
markmcclainyou'll then want to change the external network range on your setup so that the gateway on the router is on the 192.168.10/0 net17:47
*** arosen has joined #openstack-neutron17:47
*** thuc has quit IRC17:48
*** jpich has quit IRC17:48
*** thuc has joined #openstack-neutron17:48
openstackgerritA change was merged to openstack/python-neutronclient: Empty file shouldn't contain copyright nor license  https://review.openstack.org/7171417:48
*** arosen is now known as arosen317:50
*** arosen3 has quit IRC17:52
*** thuc has quit IRC17:52
*** zhipeng1 has joined #openstack-neutron17:53
*** zhipeng has quit IRC17:54
*** thuc has joined #openstack-neutron17:56
*** zhipeng1 has quit IRC17:57
*** networkstatic has quit IRC17:59
*** spandhe has joined #openstack-neutron18:00
*** jorgem1 is now known as jorgem18:00
*** zhipeng has joined #openstack-neutron18:02
*** dvorkinista has quit IRC18:02
*** claudiub___ has quit IRC18:03
*** zhipeng has quit IRC18:04
*** dvorkinista has joined #openstack-neutron18:04
Guest58201markmcclain: ok then on the topology, the blue network ext_net, i need to change the range to 192.168.0.10 (my network on eth2)18:05
*** zhipeng has joined #openstack-neutron18:05
*** marun has joined #openstack-neutron18:05
Guest58201and plug in (?) the br-ex on eth2 ?18:06
Guest58201sorry for my english :|18:06
*** dvorkinista has quit IRC18:07
*** dvorkinista has joined #openstack-neutron18:08
*** RajeshMohan has quit IRC18:13
openstackgerritJustin Hammond proposed a change to openstack/python-neutronclient: Adds support for os-auth-strategy=noauth  https://review.openstack.org/5751718:14
*** RajeshMohan has joined #openstack-neutron18:14
*** jp_at_hp1 has quit IRC18:14
*** salv-orlando has joined #openstack-neutron18:15
salv-orlandoaloha folks. I am pondering the efficiency of the two patches we've attempted to revert re: bug 128352218:16
salv-orlandoaka "the lock wait timeout thing"18:16
*** sweston has joined #openstack-neutron18:16
salv-orlandorevert #1: https://review.openstack.org/#/c/80354/ showed 5 lock wait timeout episodes (none of them caused a nova timeout)18:17
salv-orlandoall these 5 events are of the same nature of the lock wait timeout issues affecting the gate18:18
*** Guest58201 has quit IRC18:19
*** networkstatic has joined #openstack-neutron18:20
salv-orlandothe 5 events are actually only 3 error instances (in some cases both deadlocked transactions time out)18:20
*** harmw_ is now known as harmw18:20
salv-orlandorevert #2: revert #2: https://review.openstack.org/#/c/80356/  showed 2 lock wait timeouts18:20
salv-orlandohowever, the nature here is different from the one observed on the gate, which always involve a delete_port call18:20
salv-orlandoalso, the 2 events represent only a single error instance18:21
*** dvorkini_ has joined #openstack-neutron18:21
salv-orlandoso far it seems that revert #2 might be more efficient than revert #1 (which perhaps has no benefit at all)18:21
salv-orlandobut since the patch we're reverting did not do anything wrong, I'd conclude that the only issue with this patch is that it made an existing problem worse18:22
salv-orlandoso I would invite you at looking at a potential simple solution, discussed with dansmith and comstud: https://review.openstack.org/#/c/8041318:23
salv-orlandoand weigh on the comments other reviewers already left on the patch18:23
salv-orlandoI'm not sure if anybody is reading here, perhaps I should send a message to the ml?18:24
*** dvorkinista has quit IRC18:24
openstackgerritA change was merged to openstack/neutron: Fix in admin_state_up check function  https://review.openstack.org/7982218:28
openstackgerritA change was merged to openstack/python-neutronclient: Use six.moves.urllib.parse to replace urlparse  https://review.openstack.org/7350918:28
kevinbentonsalv-orlando: is there not still risk there?18:28
kevinbentondelete_network will cleanup ports18:29
kevinbentonor delete_subnet18:29
kevinbentonrather18:29
*** rwsu has quit IRC18:30
kevinbentonhttps://github.com/bigswitch/neutron/blob/master/neutron/db/db_base_plugin_v2.py#L121718:30
kevinbentonsalv-orlando: ^^18:30
openstackgerritA change was merged to openstack/python-neutronclient: Enable to select specific network service type  https://review.openstack.org/5453418:31
*** jgrimm has joined #openstack-neutron18:31
salv-orlandokevinbenton: you also need to add to the list delete_floatingip and delete_vip18:36
salv-orlandoprobably there will be more.18:36
salv-orlandothey all call delete_port, and then the lock will kick in. Now the problem might occur if they do so from within a transaction.18:37
kevinbentonsalv-orlando: yeah, I'm trying to resolve the same issues in the big switch plugin where the problem is amplified by the rest call in the transaction which causes eventlet to yield18:37
salv-orlandoI'm keeping the patch spinning through checks18:37
kevinbentondelete_subnet doesn't call delete_port through, right?18:38
salv-orlandobut if anybody has a better idea, feel free to down vote it. I am not yet proposing it as a solution. Just a test so far.18:38
*** salv-orlando has quit IRC18:38
*** salv-orlando has joined #openstack-neutron18:38
*** seagulls is now known as seagulls_coffee18:38
*** bjornar has quit IRC18:49
*** dave_tucker is now known as dave_tucker_zzz18:52
*** blogan has joined #openstack-neutron18:53
*** leseb has joined #openstack-neutron18:53
salv-orlandoa few numbers from the past 24 hours… gate job failures 702. success 9,455 (failure ratio 6.9%)18:57
*** leseb has quit IRC18:58
salv-orlandoneutron failures:18:58
salv-orlandogate-tempest-dsvm-neutro-large-ops 4.69% (33)18:58
salv-orlandogate-tempest-dsvm-neutron 26 (3.69%)18:59
*** thuc has quit IRC18:59
*** seagulls_coffee is now known as seagulls18:59
*** thuc has joined #openstack-neutron18:59
salv-orlandounit test failures: 4.12% (29)18:59
salv-orlandooverall neutron failures in the gate: 12.5%19:00
salv-orlandooverall failure ratio due to neutron: 0.87%19:01
*** salv-orlando has quit IRC19:02
*** tstevenson has joined #openstack-neutron19:02
*** jogo is now known as flashgordon19:03
*** thuc has quit IRC19:04
*** jistr has joined #openstack-neutron19:04
*** RajeshMohan has quit IRC19:04
*** RajeshMohan has joined #openstack-neutron19:06
*** thuc has joined #openstack-neutron19:07
*** baoli has quit IRC19:07
*** tstevenson has quit IRC19:08
*** jprovazn is now known as jprovazn_afk19:09
*** baoli has joined #openstack-neutron19:09
*** marun has quit IRC19:11
openstackgerritenikanorov proposed a change to openstack/neutron: LBaaS: Fix discrepancy between status and admin_state_up  https://review.openstack.org/5503219:17
*** baoli has quit IRC19:18
markmcclainsalv-orlando: Im guessing the gate failures are before we added the skip to bad unit test19:18
*** WackoRobie has quit IRC19:19
*** sfox1 has quit IRC19:20
kevinbentonwhy is this locked in update mode? https://github.com/bigswitch/neutron/blob/master/neutron/db/db_base_plugin_v2.py#L136119:21
*** manishg has joined #openstack-neutron19:23
kevinbentonit could be made safer from the evetlet/mysql deadlocks if it didn't lock in update and ignored port doesn't exist exceptions19:24
*** zzelle has joined #openstack-neutron19:27
markmcclainkevinbenton: interesting19:28
markmcclainarmax: ^^19:28
kevinbentonmarkmcclain: worth posting a patch to start a discussion?19:28
markmcclainarmax : ^^^19:28
*** zhipeng has quit IRC19:28
*** zhipeng has joined #openstack-neutron19:29
*** bjornar has joined #openstack-neutron19:29
markmcclainI think so19:29
markmcclainso delete ports is only called via rpc call from dhcp agent19:29
markmcclainso we can look at fixing that interaction19:29
kevinbentonok19:30
markmcclainarmax worked on that code last fall so he might have ideas too19:30
*** nijaba has quit IRC19:33
*** sweston has quit IRC19:38
*** nijaba has joined #openstack-neutron19:38
*** nijaba has quit IRC19:38
*** nijaba has joined #openstack-neutron19:38
*** nijaba has quit IRC19:40
*** spandhe has quit IRC19:41
*** spandhe has joined #openstack-neutron19:43
openstackgerritKevin Benton proposed a change to openstack/neutron: Don't use update lockmode on delete_ports  https://review.openstack.org/8067019:44
kevinbentonmarkmcclain, armax, salv-orlando: ^^19:44
*** carl_baldwin_ has joined #openstack-neutron19:46
*** carl_baldwin has quit IRC19:46
*** carl_baldwin_ is now known as carl_baldwin19:46
*** leseb has joined #openstack-neutron19:47
openstackgerritRobert Kukura proposed a change to openstack/neutron: ML2: Remove validate_port_binding() and unbind_port()  https://review.openstack.org/7951119:52
*** nijaba has joined #openstack-neutron19:52
*** spandhe has quit IRC19:55
*** otherwiseguy has quit IRC19:56
*** sweston has joined #openstack-neutron20:00
*** bjornar has quit IRC20:00
*** arosen has joined #openstack-neutron20:01
*** jistr has quit IRC20:05
*** sweston has quit IRC20:06
*** sweston has joined #openstack-neutron20:07
*** yamahata has quit IRC20:07
*** geekinutah has quit IRC20:08
*** geekinutah has joined #openstack-neutron20:09
*** pcm_ has quit IRC20:10
openstackgerritKevin Benton proposed a change to openstack/neutron: BigSwitch: Sync workaround for port del deadlock  https://review.openstack.org/8068820:17
openstackgerritKevin Benton proposed a change to openstack/neutron: BigSwitch: Sync workaround for port del deadlock  https://review.openstack.org/8068820:20
openstackgerritA change was merged to openstack/neutron: NSX: remove last of unneed quantum references  https://review.openstack.org/8038820:21
*** marun has joined #openstack-neutron20:25
armaxmarkmclain: back20:25
openstackgerritSridar Kandaswamy proposed a change to openstack/neutron: Fix KeyError except on router_info in FW Agent  https://review.openstack.org/7925420:27
*** rwsu has joined #openstack-neutron20:28
*** ajo has quit IRC20:29
*** jorgem has quit IRC20:30
*** otherwiseguy has joined #openstack-neutron20:36
*** jorgem has joined #openstack-neutron20:36
*** peristeri has quit IRC20:38
*** WackoRobie has joined #openstack-neutron20:40
*** spandhe has joined #openstack-neutron20:42
bloganwhat is the proper way to do a code change that changes few lines in a file but across many files? through one single commit or through many commits?20:43
*** julim has quit IRC20:47
*** bearhands has joined #openstack-neutron20:49
*** rkukura has quit IRC20:49
*** jprovazn_afk has quit IRC20:49
damnsmithbearhands: marun has a question about things you know about20:49
bearhandsok20:49
damnsmithbearhands: https://review.openstack.org/#/c/22276/20:49
bearhandsyeah20:50
bearhandsthat has changed slightly since then20:50
marunso, i'm not sure I understand why a deadlock is something that shouldn't be fixed in the code rather than retried20:50
*** rkukura has joined #openstack-neutron20:51
bearhandsideally it should be fixed20:51
bearhandsbut that's easier said than done20:51
marunwhy, what's the fix?20:51
bearhandsnot doing a SELECT and then an UPDATE20:52
marunthe original nova bug refers to performing insert and update in the same transaction20:52
marunthe targeted neutron bug would appear to be a different story20:52
bearhandsthat may be another case20:52
maruni.e. deleting lots of ports20:52
bearhandsbut SELECT and UPDATE can also do it20:52
maruni don't get how select and update have anything to do with port deletion20:52
bearhandsi'm not relating it to port deletion20:52
maruni'm seriously considering -2'ing the patch because the fix doesn't match the bug20:52
markmcclainblogan: depends on what you're changing20:53
bearhandsi'm just stating another condition that can cause it20:53
bearhands(deadlock)20:53
bearhands(or a lock wait timeout)20:53
marunhttps://review.openstack.org/#/c/78643/20:53
marun^ neutron patch in question20:53
*** vkozhukalov_ has quit IRC20:53
*** carl_baldwin has quit IRC20:54
*** nszceta_ has joined #openstack-neutron20:54
bearhands_delete_port seems to do a select and then an update20:54
marunbearhands: I'm worried that the neutron patch in question is trying to solve a deadlock bug that doesn't appear to be related to the nova bug whose patches are being cherrypicked.  Do you think my concerns are misplaced?20:54
bloganmarkmclain: its enabling the H302 rule (only import modules), i was originally going to do it one bug report at a time, but someone suggested to do it all at once, or do large portions at once20:54
bearhandsmarun: sec, brb20:55
marunbearhands: hmmm, I guess I need to look deeper.  Requiring that for deletion isn't clicking for me.20:55
damnsmithmarun: there's a LOG inside the transaction20:56
bloganmarkmcclain: thought i'd do something that was easy to learn the whole process but since doing it for all of neutron will end up being a large change im not sure what the best way to do it is20:56
marundamnsmith: we're doing cascading deletion in code :(20:56
bearhandsmarun: in general, I don't think your concern is misplaced20:56
*** zhipeng has quit IRC20:57
bearhandsmarun: if the deadlock can be avoided, obviously you should try to fix that. But it does appear that neutron is doing similar things that nova is doing...20:57
maruni.e. doing stupid things20:57
bearhandsmarun: which required us to add the decorator (for now, at least)20:57
marun*sigh*20:57
marunok20:57
bearhandswell, possibly. :)20:57
*** zhipeng has joined #openstack-neutron20:57
bearhandsit's not completely straight forward that it's stupid20:57
bearhandsi mean, completely obvious20:58
bearhandsbut there are a lot of places where we should maybe just UPDATE .. WHERE ..20:58
marunthe fact that we are doing cascade deletion in code is a pretty obvious case of stupid20:59
maruni'm not sure why openstack is so allergic to triggers20:59
*** zhipeng has quit IRC20:59
*** zhipeng has joined #openstack-neutron21:00
*** jp_at_hp has joined #openstack-neutron21:00
markmcclainblogan: for first change I'd pick something really small and with limited files/lines, so that the process is the focus21:00
markmcclainblogan: yeah hacking compliance changes should be done as one unit otherwise it is possible changes will land in interim that conflict21:01
*** salv-orlando has joined #openstack-neutron21:02
*** salv-orlando has quit IRC21:02
*** salv-orlando has joined #openstack-neutron21:02
bloganmarkmcclain: i agree and going through all the changes that would have to happen in one unit and the chain effect of other pep8 violations resulting led me to believe this was too much for a first time commit21:04
bloganmarkmcclain: i'll backlog that item and do a smaller one as I originally intended and revisit this one when i get more experience21:06
markmcclainblogan: if you wanting to learn process.. this should be easy change https://bugs.launchpad.net/neutron/+bug/121527021:07
markmcclainand it is limited to 1 file21:07
bloganmarkmcclain: thanks! i will certainly do that one21:08
bloganmarkmcclain: for future reference though, would a large change like that one be best done in one commit or broken up into logical commits?21:09
*** nszceta_ is now known as nszceta21:09
markmcclainhacking like that should be done in 1 commit21:09
markmcclainotherwise interim code is likely to land which you'll keep chasing21:10
*** leseb has quit IRC21:10
*** sabari has joined #openstack-neutron21:12
bloganah maybe i misunderstood that part of the process, i assumed you could push a branch to gerrit with many commits21:13
*** WackoRobie has quit IRC21:13
marunbearhands, damnsmith: I've found a wtf already in _port_delete...21:14
*** WackoRobie has joined #openstack-neutron21:14
marunbearhands, damnsmith: https://github.com/openstack/neutron/blob/master/neutron/db/db_base_plugin_v2.py#L137221:14
marunbearhands, damnsmith: https://github.com/openstack/neutron/blob/master/neutron/db/models_v2.py#L9321:15
marunCan either of you see a reason to be manually deleting ip allocations when a perfectly good cascade delete trigger is already defined?21:15
damnsmithmarun: is pool an additional consideration? I can't tell what's being done differently there though21:17
damnsmithseems like both the if and else clauses are doing the same thing?21:17
bearhandsmarun: I don't know that stuff well enough to say21:17
*** thedodd has quit IRC21:17
damnsmithoh, no21:17
marundamnsmith: not that I can see21:18
damnsmithwait, yeah, the wrapping was messing me up21:18
damnsmithhrm21:18
*** WackoRobie has quit IRC21:18
damnsmithmarun: yeah, I dunno, seems like the manual allocation delete is unnecessary21:18
marunI mean, apparently it's important to log when an ip won't be recycled.  But that shouldn't require a different deletion method.21:18
damnsmithheh21:18
damnsmithright21:19
marunmarkmcclain: I think we need your opinion ^^21:20
*** openstackgerrit has quit IRC21:21
*** openstackgerrit has joined #openstack-neutron21:21
marunmarkmcclain: it appears as if there should be no need to do anything but delete the port21:22
*** sfox has joined #openstack-neutron21:22
marunmarkmcclain: which would remove the locks and maybe remove the possibility of deadlock21:22
marunmarkmcclain: but maybe I'm missing something21:22
*** zhipeng has quit IRC21:24
markmcclainblogan: you can push many commits, but ideally the reviews should be a single logical unit so that history stays clean and reverts are easy should it be needed21:25
*** zhipeng has joined #openstack-neutron21:26
markmcclainmarun: we were having issues with duplicate IPs21:27
marunmarkmcclain: so it's to paper over a deeper bug?21:27
marunmarkmcclain: wait… that still doesn't make sense to me...21:27
marunmarkmcclain: There is a relationship between ipallocation and port with a cascade delete21:27
marunmarkmcclain: why is it not sufficient to delete the port and why does ip address even come into it?21:28
markmcclainwell that's because this is a left over21:28
markmcclainlook at havana21:28
markmcclainhttps://github.com/openstack/neutron/blob/stable/havana/neutron/db/db_base_plugin_v2.py#L141221:28
*** stiggie has joined #openstack-neutron21:28
markmcclainactually wait the code is the same21:29
marunmarkmcclain: so in what case would the current model definition not obviate the need for manual removal of ipallocation?21:29
marunmarkmcclain: I'm guessing that something changed, I'm just wanting to see if we can clean this up and not have to rely on a retry_on_deadlock decorator21:29
markmcclainok.. so there is a subtle difference21:30
markmcclainpreviously we would recycle the ip and compact the pool on the spot21:30
markmcclainso locking the table was necessary21:31
markmcclainnow with _delete_ip being called we're just manually deleting it21:31
*** openstackgerrit has quit IRC21:32
*** openstackgerrit has joined #openstack-neutron21:32
*** dvorkini_ has quit IRC21:32
markmcclainmarun: here's the bad part21:33
markmcclainhttps://github.com/openstack/neutron/blob/master/neutron/db/db_base_plugin_v2.py#L137921:33
marunmarkmcclain: why is that necessary?21:34
markmcclainthis is conditional runs another query and has _no_ impact because the logic is the same for if and else branch21:34
marunmarkmcclain: right.21:34
*** dims has quit IRC21:34
markmcclainso sadly we're still going to need to call NeutronDbPluginV2._delete_ip_allocation()21:35
marunmarkmcclain: again, since the ipallocation is queried by port id, cascade delete will ensure deletion of every ipallocation being examined in the current code21:35
marunmarkmcclain: here's where you explain why we have a cascade deletion then...21:35
markmcclainbecause someone could have done some really funny overriding21:35
marunmarkmcclain: can't we search pretty easily for that?21:36
markmcclaingranted inheriting and then overriding a static method on the base class basically requires monkeypatching21:36
*** dvorkinista has joined #openstack-neutron21:37
*** vkozhukalov_ has joined #openstack-neutron21:37
marunmarkmcclain: if it's being done in tree, we can address it21:38
markmcclainyeah21:38
markmcclainalso this will save a subnet lookup too21:38
marunmarkmcclain: + we have tests ;)21:38
marunmarkmcclain: ok, I'm going to -2 the retry_on_deadlock patch pending a patch that will clean up _port_delete.  We may still need it, but hopefully not.21:39
*** blogan_ has joined #openstack-neutron21:39
*** blogan has quit IRC21:42
openstackgerritA change was merged to openstack/neutron: Send fdb remove message when a port is migrated  https://review.openstack.org/6176721:43
*** pba has quit IRC21:47
*** pba has joined #openstack-neutron21:47
*** jp_at_hp has quit IRC21:48
*** vjay has quit IRC21:48
*** suresh12 has quit IRC21:48
*** zhipeng has quit IRC21:50
*** zhipeng has joined #openstack-neutron21:51
*** zhipeng has quit IRC21:52
*** dims has joined #openstack-neutron21:53
*** zhipeng has joined #openstack-neutron21:53
*** nszceta has quit IRC21:54
*** changbl has quit IRC21:55
*** bada_ has joined #openstack-neutron21:55
*** arosen has quit IRC21:55
*** changbl has joined #openstack-neutron21:55
*** dvorkinista has quit IRC21:55
*** bada__ has quit IRC21:56
*** nszceta has joined #openstack-neutron21:57
*** jobewan has quit IRC21:58
*** stiggie has quit IRC22:02
openstackgerritA change was merged to openstack/neutron: Imported Translations from Transifex  https://review.openstack.org/8049222:03
*** zhipeng has quit IRC22:10
*** sacharya has quit IRC22:14
*** zhipeng has joined #openstack-neutron22:14
*** zhipeng1 has joined #openstack-neutron22:17
*** banix has quit IRC22:18
*** zhipeng has quit IRC22:19
openstackgerritKevin Benton proposed a change to openstack/neutron: BigSwitch: Sync workaround for port del deadlock  https://review.openstack.org/8068822:20
*** networkstatic is now known as networkstatic_zZ22:20
*** leseb has joined #openstack-neutron22:21
*** zhipeng has joined #openstack-neutron22:24
*** zarric_ has joined #openstack-neutron22:26
*** arosen has joined #openstack-neutron22:26
*** zhipeng1 has quit IRC22:26
*** manishg has quit IRC22:26
*** leseb has quit IRC22:26
*** zarric has quit IRC22:26
*** zarric_ is now known as zarric22:26
*** ihrachys has quit IRC22:26
*** manishg has joined #openstack-neutron22:32
*** arosen has quit IRC22:33
*** sfox has quit IRC22:33
*** zhipeng has quit IRC22:33
*** arosen has joined #openstack-neutron22:34
*** zhipeng has joined #openstack-neutron22:34
*** manishg has quit IRC22:37
*** manishg has joined #openstack-neutron22:38
openstackgerritArmando Migliaccio proposed a change to openstack/neutron: Fix misleading error message about failed dhcp notifications  https://review.openstack.org/7945822:40
*** marun has quit IRC22:42
*** armax has quit IRC22:42
*** sfox has joined #openstack-neutron22:42
*** zhipeng has quit IRC22:43
*** zhipeng has joined #openstack-neutron22:44
*** jorgem has quit IRC22:45
*** manishg has quit IRC22:47
*** manishg has joined #openstack-neutron22:49
*** zhipeng has quit IRC22:49
*** zhipeng has joined #openstack-neutron22:49
*** manishg has quit IRC22:50
*** arosen has quit IRC22:55
*** lucas__ has joined #openstack-neutron22:56
*** luqas has quit IRC22:58
*** alagalah has quit IRC22:58
openstackgerritAkihiro Motoki proposed a change to openstack/neutron: Ensure to count firewalls in target tenant  https://review.openstack.org/8071523:00
openstackgerritAkihiro Motoki proposed a change to openstack/neutron: NEC plugin: Rename quantum_id column to neutron_id  https://review.openstack.org/7885423:03
*** rohit404 has quit IRC23:06
*** dvorkinista has joined #openstack-neutron23:06
openstackgerritDhanashree Gosavi proposed a change to openstack/neutron: Validate multicast ip range in Cisco N1kv Plugin  https://review.openstack.org/7529123:07
*** dvorkinista has quit IRC23:11
*** zhipeng has quit IRC23:13
*** zhipeng has joined #openstack-neutron23:14
*** rohit404 has joined #openstack-neutron23:14
*** ywu has quit IRC23:18
*** ywu has joined #openstack-neutron23:18
*** ijw has quit IRC23:19
*** sweston has quit IRC23:27
*** djoreilly has quit IRC23:30
*** otherwiseguy has quit IRC23:32
*** openstackstatus has quit IRC23:36
*** rwsu has quit IRC23:38
*** openstackgerrit has quit IRC23:39
*** rossella_s has quit IRC23:39
*** rm_work has quit IRC23:39
*** changbl has quit IRC23:42
*** nati_ueno has quit IRC23:42
*** HenryG has quit IRC23:42
*** thuc has quit IRC23:43
*** mwagner_lap has quit IRC23:45
*** zhipeng has quit IRC23:45
*** yongli has quit IRC23:46
*** joshh has quit IRC23:46
*** seagulls has quit IRC23:46
*** SpamapS has quit IRC23:46
*** sdague has quit IRC23:46
*** lucas__ has quit IRC23:46
*** nszceta has quit IRC23:46
*** harmw has quit IRC23:46
*** aryan has quit IRC23:46
*** mtreinish has quit IRC23:46
*** sc68cal has quit IRC23:46
*** eezhova has quit IRC23:46
*** ToM has quit IRC23:46
*** bada_ has quit IRC23:46
*** vkozhukalov_ has quit IRC23:46
*** harlowja has quit IRC23:46
*** fouxm has quit IRC23:46
*** damnsmith has quit IRC23:46
*** enikanorov has quit IRC23:46
*** feleouet has quit IRC23:46
*** Matt2 has quit IRC23:46
*** asadoughi has quit IRC23:46
*** dgrig has quit IRC23:46
*** ywu has quit IRC23:46
*** rohit404 has quit IRC23:46
*** zarric has quit IRC23:46
*** pba has quit IRC23:46
*** blogan_ has quit IRC23:46
*** sbalukoff has quit IRC23:46
*** EmilienM has quit IRC23:46
*** obondarev has quit IRC23:46
*** markvoelker has quit IRC23:46
*** lucian1 has quit IRC23:46
*** mancdaz has quit IRC23:46
*** gizmoguy has quit IRC23:46
*** edhall has quit IRC23:46
*** salv-orlando has quit IRC23:46
*** rkukura has quit IRC23:46
*** zzelle has quit IRC23:46
*** jgrimm has quit IRC23:46
*** crc32 has quit IRC23:46
*** jecarey has quit IRC23:46
*** sabari has quit IRC23:46
*** JoeJulian has quit IRC23:46
*** chrisf has quit IRC23:46
*** ptoohill has quit IRC23:46
*** doude has quit IRC23:46
*** ilyashakhat has quit IRC23:46
*** sbadia has quit IRC23:46
*** dkehn has quit IRC23:46
*** pradipta_away has quit IRC23:46
*** _cerberus_ has quit IRC23:46
*** morgabra has quit IRC23:46
*** kevinbenton has quit IRC23:46
*** spandhe has quit IRC23:46
*** RajeshMohan has quit IRC23:46
*** networkstatic_zZ has quit IRC23:46
*** SumitNaiksatam has quit IRC23:46
*** _cjones_ has quit IRC23:46
*** markwash has quit IRC23:46
*** alexpilotti has quit IRC23:46
*** yfried has quit IRC23:46
*** pcarver has quit IRC23:46
*** JoeHazzers has quit IRC23:46
*** thurloat has quit IRC23:46
*** juice has quit IRC23:46
*** jaypipes has quit IRC23:46
*** cgoncalves has quit IRC23:46
*** yamahata_ has quit IRC23:46
*** Qlawy has quit IRC23:46
*** annashen_ has quit IRC23:46
*** rha has quit IRC23:46
*** linuxaddicts has quit IRC23:46
*** geekinutah has quit IRC23:46
*** nijaba has quit IRC23:46
*** emagana has quit IRC23:46
*** enikanorov_ has quit IRC23:46
*** jianingy_afk has quit IRC23:46
*** rdo has quit IRC23:46
*** skath has quit IRC23:46
*** wayneeseguin has quit IRC23:46
*** skraynev_afk has quit IRC23:46
*** dkehn_ has quit IRC23:46
*** andyhky has quit IRC23:46
*** sgran has quit IRC23:46
*** sff has quit IRC23:46
*** inara` has quit IRC23:46
*** haleyb has quit IRC23:46
*** bearhands has quit IRC23:46
*** dosaboy has quit IRC23:46
*** flashgordon has quit IRC23:46
*** roaet has quit IRC23:46
*** qiuyu has quit IRC23:46
*** lifeless has quit IRC23:46
*** marios has quit IRC23:46
*** mordred has quit IRC23:46
*** cburgess has quit IRC23:46
*** Dseven has quit IRC23:46
*** zigo has quit IRC23:46
*** nplanel has quit IRC23:46
*** vishy has quit IRC23:46
*** puck has quit IRC23:46
*** openstack has joined #openstack-neutron23:47
*** openstackstatus has joined #openstack-neutron23:47
*** aryan has joined #openstack-neutron23:50
*** bada_ has joined #openstack-neutron23:54

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