| *** kaisers_ has joined #openstack-manila | 00:03 | |
| *** kaisers_ has quit IRC | 00:08 | |
| *** furlongm_ has quit IRC | 00:19 | |
| *** Suyi_ has quit IRC | 00:19 | |
| *** qeelee has joined #openstack-manila | 00:33 | |
| *** mtanino has quit IRC | 00:44 | |
| *** gouthamr_ has quit IRC | 00:47 | |
| *** yangyapeng has joined #openstack-manila | 00:52 | |
| *** esker has quit IRC | 00:55 | |
| *** chlong has joined #openstack-manila | 01:18 | |
| *** cknight has joined #openstack-manila | 01:23 | |
| *** qeelee has quit IRC | 01:27 | |
| openstackgerrit | Yingzhe Zeng proposed openstack/manila: Implement replication support in huawei driver https://review.openstack.org/331586 | 01:30 |
|---|---|---|
| *** wangqun has joined #openstack-manila | 01:33 | |
| *** cknight has quit IRC | 01:33 | |
| *** qeelee has joined #openstack-manila | 01:37 | |
| *** kaisers_ has joined #openstack-manila | 01:53 | |
| *** kaisers_ has quit IRC | 01:57 | |
| *** xiaoyang has joined #openstack-manila | 02:11 | |
| xiaoyang | hello,manila | 02:13 |
| *** qeelee has quit IRC | 02:16 | |
| xiaoyang | how to use L2 to provide share server? | 02:17 |
| *** qeelee has joined #openstack-manila | 02:17 | |
| *** vbellur has joined #openstack-manila | 02:18 | |
| xiaoyang | 有人吗？ | 02:47 |
| *** ganso has quit IRC | 02:51 | |
| *** alkhodos has quit IRC | 03:04 | |
| *** qeelee has quit IRC | 03:06 | |
| *** qeelee has joined #openstack-manila | 03:12 | |
| *** dustins has joined #openstack-manila | 03:28 | |
| *** zhangyi has joined #openstack-manila | 03:28 | |
| *** dustins has quit IRC | 03:37 | |
| *** chlong has quit IRC | 03:37 | |
| *** dustins_ has joined #openstack-manila | 03:41 | |
| *** kaisers_ has joined #openstack-manila | 03:41 | |
| *** dustins has joined #openstack-manila | 03:42 | |
| *** dustins_ has quit IRC | 03:44 | |
| *** kaisers_ has quit IRC | 03:46 | |
| *** dustins has quit IRC | 03:59 | |
| *** chlong has joined #openstack-manila | 04:10 | |
| *** daidv has joined #openstack-manila | 04:31 | |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/manila: Updated from global requirements https://review.openstack.org/360891 | 05:08 |
| *** chlong has quit IRC | 05:21 | |
| *** kaisers_ has joined #openstack-manila | 05:30 | |
| *** sandanar has joined #openstack-manila | 05:34 | |
| *** kaisers_ has quit IRC | 05:35 | |
| *** daidv has quit IRC | 05:37 | |
| *** daidv_ has joined #openstack-manila | 05:37 | |
| *** nkrinner_afk is now known as nkrinner | 05:43 | |
| *** gouthamr has joined #openstack-manila | 05:44 | |
| *** lpetrut has joined #openstack-manila | 06:29 | |
| openstackgerrit | Yingzhe Zeng proposed openstack/manila: huawei driver default create thin type share https://review.openstack.org/351475 | 06:51 |
| *** aovchinnikov has joined #openstack-manila | 06:51 | |
| *** kaisers_ has joined #openstack-manila | 07:19 | |
| *** kaisers_ has quit IRC | 07:24 | |
| openstackgerrit | Yingzhe Zeng proposed openstack/manila: Implement replication support in huawei driver https://review.openstack.org/331586 | 07:29 |
| *** lpetrut has quit IRC | 07:30 | |
| *** kaisers_ has joined #openstack-manila | 07:30 | |
| openstackgerrit | Alexey Ovchinnikov proposed openstack/manila: Clarify grenade failure message https://review.openstack.org/360942 | 07:34 |
| openstackgerrit | NidhiMittalHada proposed openstack/manila: manila does not calculate allocated_capacity_gb https://review.openstack.org/351034 | 07:41 |
| *** nidhimittalhada has joined #openstack-manila | 07:42 | |
| nidhimittalhada | @vponomaryov, @gouthamr @xyang @ganso @zhongjun https://review.openstack.org/315012 https://review.openstack.org/348160 Can you please review ? | 07:43 |
| *** nidhimittalhada has quit IRC | 07:43 | |
| openstackgerrit | Nam Nguyen Hoai proposed openstack/manila: TrivialFix: Remove cfg import unused https://review.openstack.org/360960 | 08:03 |
| *** zhangyi has quit IRC | 08:06 | |
| *** zhangyi has joined #openstack-manila | 08:07 | |
| *** ociuhandu has joined #openstack-manila | 08:13 | |
| *** openstackgerrit has quit IRC | 08:18 | |
| *** ociuhandu has quit IRC | 08:18 | |
| *** openstackgerrit has joined #openstack-manila | 08:19 | |
| openstackgerrit | Alexey Ovchinnikov proposed openstack/manila: [Do Not Merge] Test networking issues in tests https://review.openstack.org/360973 | 08:25 |
| *** gouthamr has quit IRC | 09:05 | |
| *** darrenc has quit IRC | 09:06 | |
| *** darrenc has joined #openstack-manila | 09:06 | |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/manila: Updated from global requirements https://review.openstack.org/360891 | 09:12 |
| *** lpetrut has joined #openstack-manila | 09:13 | |
| *** esker has joined #openstack-manila | 09:37 | |
| *** esker is now known as esker[away] | 09:37 | |
| openstackgerrit | Yingzhe Zeng proposed openstack/manila: Change test bugs for replication CI https://review.openstack.org/361042 | 09:46 |
| openstackgerrit | Cao Xuan Hoang proposed openstack/manila: Clean imports in code https://review.openstack.org/361043 | 09:48 |
| openstackgerrit | Alexey Ovchinnikov proposed openstack/manila: [Do Not Merge] More tests for generic networking behaviour https://review.openstack.org/361045 | 09:50 |
| *** esker[away] is now known as esker | 10:03 | |
| *** ociuhandu has joined #openstack-manila | 10:08 | |
| *** esker has quit IRC | 10:11 | |
| *** yangyapeng has quit IRC | 10:13 | |
| openstackgerrit | NidhiMittalHada proposed openstack/manila: manila does not calculate allocated_capacity_gb https://review.openstack.org/351034 | 10:18 |
| *** xiaoyang has quit IRC | 10:20 | |
| openstackgerrit | Nam Nguyen Hoai proposed openstack/manila: TrivialFix: Remove cfg import unused https://review.openstack.org/360960 | 10:32 |
| *** wangqun has quit IRC | 10:53 | |
| *** ganso has joined #openstack-manila | 10:58 | |
| openstackgerrit | Alexey Ovchinnikov proposed openstack/manila: Container driver: Fix for guest mounting issue https://review.openstack.org/361080 | 10:59 |
| *** JoseMello has joined #openstack-manila | 11:01 | |
| *** alyson_ has joined #openstack-manila | 11:15 | |
| openstackgerrit | zzxwill proposed openstack/manila: Put all imports from manila.i18n in one line https://review.openstack.org/361099 | 11:26 |
| *** nkrinner is now known as nkrinner_afk | 11:28 | |
| *** timcl has joined #openstack-manila | 11:40 | |
| *** akapil has joined #openstack-manila | 11:59 | |
| *** nkrinner_afk is now known as nkrinner | 12:06 | |
| *** timcl has quit IRC | 12:12 | |
| *** timcl has joined #openstack-manila | 12:13 | |
| *** akapil has quit IRC | 12:13 | |
| *** akapil has joined #openstack-manila | 12:14 | |
| *** xyang1 has joined #openstack-manila | 12:17 | |
| *** kaisers1 has quit IRC | 12:23 | |
| *** kaisers has quit IRC | 12:23 | |
| *** akapil has quit IRC | 12:25 | |
| *** kaisers has joined #openstack-manila | 12:29 | |
| *** kaisers1 has joined #openstack-manila | 12:30 | |
| *** zhangyi has quit IRC | 12:34 | |
| *** akapil has joined #openstack-manila | 12:48 | |
| *** eharney has joined #openstack-manila | 13:04 | |
| *** akapil has quit IRC | 13:14 | |
| *** rooneym has joined #openstack-manila | 13:22 | |
| *** daidv_ has quit IRC | 13:24 | |
| ganso | bswartz, xyang1, markstur, tbarron: Could you please take a look at https://review.openstack.org/#/c/332267/ ? Thanks in advance | 13:25 |
| openstackgerrit | Rodrigo Barbieri proposed openstack/manila: [DNM] Testing bugfix https://review.openstack.org/361180 | 13:32 |
| *** dustins has joined #openstack-manila | 13:37 | |
| *** gouthamr has joined #openstack-manila | 13:47 | |
| *** akerr has joined #openstack-manila | 13:48 | |
| *** nkrinner is now known as nkrinner_afk | 13:49 | |
| *** akapil has joined #openstack-manila | 13:50 | |
| *** sandanar has quit IRC | 13:53 | |
| *** akapil has quit IRC | 13:54 | |
| openstackgerrit | Emilien Macchi proposed openstack/puppet-manila: Prepare 9.2.0 (newton b3) https://review.openstack.org/361209 | 13:55 |
| *** aovchinnikov has quit IRC | 13:55 | |
| bswartz | ganso: how are we supposed to merge this with the gate acting so flaky? | 14:16 |
| *** mtanino has joined #openstack-manila | 14:17 | |
| ganso | bswartz: I have diagnosed the gate error and I am attempting to fix, it is unrelated. Still, if the code is good and worth 2 +2 and +W then I believe we should try to make it pass the gate rechecking so it unblocks other patches in the Share Migration tree, which are waiting to be reviewed. | 14:24 |
| ganso | bswartz: by the way, curiously, the gate only fails when the node running the job has IPv6 addresses | 14:24 |
| ganso | bswartz: only generic driver jobs fail | 14:24 |
| bswartz | ganso: thanks | 14:25 |
| bswartz | ganso: can you explain what the fix will look like yet? | 14:26 |
| bswartz | I know we can +2A your patch before the gate is fixed but the gate bug is distracting me from doing reviews | 14:27 |
| *** yangyapeng has joined #openstack-manila | 14:28 | |
| ganso | bswartz: the fault is: there is a step in generic driver where it updates the routes in the host. It list the routes, and attempts to delete out-of-date routes. It is listing, and then when trying to delete says the route is not there. At first I thought about a concurrency issue, although too frequent to be that, so I added a lock (in fact a fixed a lock | 14:28 |
| ganso | which was out of place, you can see in https://review.openstack.org/361180). | 14:28 |
| ganso | bswartz: so now, I will add more logging to see what the route output is and why it is listing something and then failing to delete as if it was not there | 14:29 |
| tbarron | ganso: so there's a race in _plug_interface_in_host but only for ipv6? (looking at https://review.openstack.org/#/c/361180) | 14:29 |
| ganso | bswartz: because simple lock solution did not work | 14:29 |
| ganso | tbarron: there is race, but that's not the real cause | 14:29 |
| bswartz | ganso: thanks I was on this same path but a bit behind you with my investigation | 14:29 |
| bswartz | It wouldn't surprise me if vponomaryov wrote this code without testing ipv6 and the presense of ipv6 screws something up | 14:30 |
| *** akapil has joined #openstack-manila | 14:30 | |
| *** cknight has joined #openstack-manila | 14:31 | |
| *** chlong has joined #openstack-manila | 14:32 | |
| *** kaisers_ has quit IRC | 14:32 | |
| *** Yogi1 has joined #openstack-manila | 14:35 | |
| *** qeelee has quit IRC | 14:40 | |
| openstackgerrit | Rodrigo Barbieri proposed openstack/manila: [DNM] Testing bugfix https://review.openstack.org/361180 | 14:41 |
| *** rhagarty has quit IRC | 14:45 | |
| *** timcl has quit IRC | 14:49 | |
| *** akapil has quit IRC | 14:55 | |
| *** kaisers1 has quit IRC | 14:56 | |
| *** kaisers1 has joined #openstack-manila | 14:58 | |
| *** timcl has joined #openstack-manila | 15:00 | |
| *** MikeG451 has quit IRC | 15:02 | |
| *** MikeG451 has joined #openstack-manila | 15:07 | |
| *** pcaruana has quit IRC | 15:33 | |
| *** timcl has quit IRC | 15:35 | |
| openstackgerrit | chen.xing proposed openstack/manila: [install] Create endpoint in one command https://review.openstack.org/361324 | 15:57 |
| *** chlong has quit IRC | 16:07 | |
| *** alkhodos has joined #openstack-manila | 16:17 | |
| *** chlong has joined #openstack-manila | 16:19 | |
| *** kaisers_ has joined #openstack-manila | 16:21 | |
| *** timcl has joined #openstack-manila | 16:22 | |
| *** kaisers_ has quit IRC | 16:26 | |
| openstackgerrit | David Moreau Simard proposed openstack/puppet-manila: Prepare 9.2.0 (newton b3) https://review.openstack.org/361209 | 16:26 |
| *** sandanar has joined #openstack-manila | 16:32 | |
| *** david-lyle has joined #openstack-manila | 16:36 | |
| *** Suyi_ has joined #openstack-manila | 16:56 | |
| openstackgerrit | Alexey Khodos proposed openstack/manila: Nexenta: adding share drivers for NexentaStor https://review.openstack.org/309286 | 16:57 |
| openstackgerrit | Alexey Khodos proposed openstack/manila: Nexenta: adding share drivers for NexentaStor https://review.openstack.org/309286 | 16:59 |
| *** chlong has quit IRC | 17:02 | |
| *** lpetrut has quit IRC | 17:07 | |
| *** aovchinnikov has joined #openstack-manila | 17:17 | |
| openstackgerrit | David Moreau Simard proposed openstack/puppet-manila: Prepare 9.2.0 (newton b3) https://review.openstack.org/361209 | 17:19 |
| openstackgerrit | David Moreau Simard proposed openstack/puppet-manila: Tag manilaclient package with 'openstack' to inherit openstacklib defaults https://review.openstack.org/361391 | 17:19 |
| *** DericHorn-HP has joined #openstack-manila | 17:26 | |
| openstackgerrit | David Moreau Simard proposed openstack/puppet-manila: Prepare 9.2.0 (newton b3) https://review.openstack.org/361209 | 17:27 |
| openstackgerrit | David Moreau Simard proposed openstack/puppet-manila: Tag manilaclient package with 'openstack' to inherit openstacklib defaults https://review.openstack.org/361391 | 17:27 |
| openstackgerrit | Alexey Khodos proposed openstack/manila: Nexenta: adding share drivers for NexentaStor https://review.openstack.org/309286 | 17:33 |
| *** alkhodos has quit IRC | 17:41 | |
| *** tpsilva has joined #openstack-manila | 18:00 | |
| *** eharney has quit IRC | 18:03 | |
| *** sandanar has quit IRC | 18:06 | |
| *** ociuhandu has quit IRC | 18:06 | |
| *** kaisers_ has joined #openstack-manila | 18:10 | |
| *** aovchinnikov has quit IRC | 18:11 | |
| *** timcl has quit IRC | 18:11 | |
| openstackgerrit | David Moreau Simard proposed openstack/puppet-manila: Prepare 9.2.0 (newton b3) https://review.openstack.org/361209 | 18:11 |
| ganso | bswartz: problem is in ip route it seems | 18:14 |
| ganso | bswartz: "ip route" command | 18:14 |
| *** kaisers_ has quit IRC | 18:15 | |
| openstackgerrit | Merged openstack/puppet-manila: Tag manilaclient package with 'openstack' to inherit openstacklib defaults https://review.openstack.org/361391 | 18:26 |
| openstackgerrit | Rodrigo Barbieri proposed openstack/manila: [CI FIX] Fix 'ip route' matching multiple subnets https://review.openstack.org/361413 | 18:30 |
| bswartz | ganso: it looked okay to me | 18:30 |
| bswartz | ganso: what exactly was wrong with it? | 18:31 |
| ganso | bswartz: 'ip route list ... match <subnet>' | 18:32 |
| ganso | bswartz: ubuntu@ubuntu:~$ ip route list proto kernel match 10.1.0.0/20 dev tap50dea090-07 | 18:32 |
| ganso | 10.0.0.0/8 scope link | 18:32 |
| ganso | 10.1.0.0/20 scope link | 18:32 |
| bswartz | oh I thought you were referring to the ' ip route del' command | 18:33 |
| ganso | bswartz: that makes no sense to me ^ | 18:33 |
| bswartz | the command that failed | 18:33 |
| bswartz | ganso: why not? | 18:33 |
| ganso | bswartz: it probably does not make sense to me because I don't expect 10.0.0.8/20 to show up if I requested 10.1.0.0/20 | 18:34 |
| ganso | bswartz: I don't know the logic and purpose of 'match' parameter | 18:34 |
| ganso | bswartz: but there is an 'exact' parameter, which can be used instead of 'match' that fixes | 18:34 |
| ganso | bswartz: I would expect 'match' to do what 'exact' does from the beginning | 18:34 |
| *** timcl has joined #openstack-manila | 18:35 | |
| *** dustins has quit IRC | 18:35 | |
| *** baojg has joined #openstack-manila | 18:35 | |
| bswartz | match returns all supernets as well | 18:36 |
| bswartz | according to the docs | 18:36 |
| ganso | bswartz: the docs probably weren't checked when the code was written | 18:38 |
| bswartz | perhaps | 18:38 |
| bswartz | it could just a bug where the author misunderstood the iproute2 command | 18:38 |
| bswartz | and it didn't end up mattering until now | 18:39 |
| ganso | bswartz: yes, seems exact like it | 18:39 |
| ganso | bswartz: took 2.5 years to blow up lol | 18:40 |
| bswartz | well it still makes me nervous because I don't understand what the purpose of the code is | 18:41 |
| ganso | bswartz: which code/purpose exactly? | 18:48 |
| ganso | bswartz: what part you don't understand? | 18:49 |
| *** baojg has quit IRC | 18:49 | |
| *** eharney has joined #openstack-manila | 18:49 | |
| *** ociuhandu has joined #openstack-manila | 18:50 | |
| bswartz | ganso: all of this routing stuff | 18:50 |
| bswartz | It's unclear why it's necessary | 18:50 |
| bswartz | networking should be handled in neutron IMO | 18:50 |
| bswartz | I'm worried this code exists because neutron wasn't usable 2 years ago and we were implementing everything ourselves | 18:50 |
| ganso | bswartz: oh I see | 18:52 |
| ganso | bswartz: maybe | 18:52 |
| ganso | bswartz: what I understand of this code is that it maintains the routes of the tap interfaces created in the host for share servers | 18:53 |
| *** baojg has joined #openstack-manila | 18:53 | |
| ganso | bswartz: every time new share servers are created for different share_networks it adds a route, and it needs to be cleaned up later when they are deleted | 18:53 |
| bswartz | it feels hacky | 18:54 |
| ganso | bswartz: also, the routes may be conflicting... so the generic driver has this vulnerability that is not handled very well, hence why the scenario jobs failed | 18:54 |
| ganso | bswartz: it sure does lol | 18:54 |
| ganso | bswartz: it may be that this fix I just submitted may have fixed scenario job as well... I am not sure yet | 18:55 |
| *** ociuhandu has quit IRC | 18:56 | |
| ganso | bswartz: is today a holiday? haven't seen much activity here today | 18:56 |
| *** baojg has quit IRC | 18:58 | |
| *** baojg has joined #openstack-manila | 19:00 | |
| *** baojg has quit IRC | 19:00 | |
| *** baojg has joined #openstack-manila | 19:01 | |
| bswartz | ganso: no it's just the day before feature freeze and people are running around trying to merge things | 19:03 |
| bswartz | personally I was out for 2 hours for a school thing but I'm back now | 19:03 |
| *** ociuhandu has joined #openstack-manila | 19:06 | |
| openstackgerrit | Rodrigo Barbieri proposed openstack/manila: [CI FIX] Fix 'ip route' matching multiple subnets https://review.openstack.org/361413 | 19:08 |
| bswartz | ganso markstur xyang1 cknight toabctl tbarron: http://lists.openstack.org/pipermail/openstack-dev/2016-August/102358.html please let me know your opinions on this FFE request and I'll reply on the ML | 19:10 |
| tbarron | so we should make a note to look at how nova adds a route on the tap interfaces created for its instances, maybe there's a higher level routine for this setup now | 19:12 |
| tbarron | i don't mean instead of the fix ganso is doing right now | 19:12 |
| tbarron | just a backlog item | 19:12 |
| ganso | tbarron: the route is not in the instance, it is in the host | 19:12 |
| ganso | tbarron: oh, for the tap interfaces | 19:12 |
| ganso | tbarron: nevermind, misread your message | 19:12 |
| tbarron | ganso: oh, yeah, sorry | 19:12 |
| tbarron | ok, not sorry :) | 19:13 |
| tbarron | but later | 19:13 |
| ganso | tbarron: lol | 19:13 |
| *** baojg has quit IRC | 19:13 | |
| *** baojg has joined #openstack-manila | 19:15 | |
| *** ociuhandu has quit IRC | 19:17 | |
| tbarron | bswartz: i personally have no problem with FFEs for isolated features that don't put the rest of the project at risk. Often was in the uncomfortable position of having to make them in cinder when I worked for a storage vendor and my perceptions of leadership there are colored by whether they were reasonable and gracious or called me names and stuff like that :) | 19:17 |
| *** dustins has joined #openstack-manila | 19:19 | |
| *** baojg has quit IRC | 19:19 | |
| cknight | tbarron: I never called you names! :-) | 19:20 |
| tbarron | cknight: :) | 19:21 |
| *** ociuhandu has joined #openstack-manila | 19:21 | |
| cknight | tbarron, bswartz, ganso, xyang1, toabctl: Regarding this FFE, I have to vote 'no'. The patch was untouched for all of July and has *never* passed Jenkins. I don't see how it can suddenly be a priority. | 19:23 |
| ganso | bswartz, xyang1, toabctl, cknight, tbarron: I noticed it had never passed jenkins before, implying that it was not completed... although incomplete patches could also pass jenkins. It seems there had not been much development in July... but I don't see that as a big problem. My vote is OK. | 19:27 |
| *** dustins has quit IRC | 19:27 | |
| bswartz | has anyone reviewed it? I don't see the +1s/+2s | 19:29 |
| bswartz | I guess we need to fix the gate and recheck this one to know if it's passing jenkins yet as-is | 19:30 |
| ganso | although I am not happy that 7 days since FPF have passed and now it seems a bit late to start taking a look at this and submitting changes... if the dev is able to fix everything in time for FF then ok | 19:30 |
| ganso | haven't taken a look myself so I don't know how close to good to be merged it is | 19:31 |
| bswartz | ganso: yeah he keeps updating the patch not in response to any feedback | 19:31 |
| bswartz | it feels like it wasn't even done as of yesterday | 19:31 |
| ganso | bswartz: lemme check the amount of lines between patches | 19:32 |
| bswartz | PS9 is a trivial removal of whitespace, but PS8 adds new code out of the blue | 19:32 |
| bswartz | PS8 was done tuesday | 19:32 |
| ganso | PS #6..9 changed 10 lines of code | 19:33 |
| markstur | bswartz, It hasn't passed jenkins yet. I didn't push Jay to make the FPF because he had been pulled off to do cinder work. I didn't realize he would be back and so close to being ready for the FPF. | 19:33 |
| *** dustins has joined #openstack-manila | 19:33 | |
| bswartz | markstur: what's your guidance here? | 19:33 |
| bswartz | if this slips it's a major blow to HPE? | 19:33 |
| bswartz | it feels low risk, but the rules are being flouted and we want to discourage that | 19:34 |
| markstur | bswartz, I think it is a shame to just miss the F_P_F by a day, but w/o a passing jenkins it is hard to say how ready it is other than Jay says it is ready and fully tested. | 19:34 |
| markstur | It is not a major blow to HPE. Nothing requires pools yet. | 19:34 |
| markstur | It is a long set-back for Jay if it is truly ready now and he has to sit on it until O | 19:35 |
| bswartz | maybe we could approve the exception and censure jay-mehta since it's a first offense | 19:36 |
| *** dustins_ has joined #openstack-manila | 19:36 | |
| bswartz | I want to be welcoming to new community members | 19:37 |
| bswartz | but I don't want to encourage bad behavior | 19:37 |
| *** dustins has quit IRC | 19:37 | |
| xyang1 | bswartz: I may have reviewed that patch once but then I saw the -2 | 19:39 |
| *** lpetrut has joined #openstack-manila | 19:41 | |
| *** porrua has quit IRC | 19:44 | |
| bswartz | xyang1: do you have an opinion on the code? or the FFE? | 19:49 |
| *** akerr has quit IRC | 19:58 | |
| *** kaisers_ has joined #openstack-manila | 19:59 | |
| cknight | bswartz: how is censuring someone 'welcoming'? | 19:59 |
| bswartz | cknight: people prefer getting warning to getting a speeding ticket | 20:00 |
| *** ganso has quit IRC | 20:00 | |
| markstur | cknight, :) | 20:00 |
| cknight | markstur: It's tough to miss FPF by a day, but revert-to-snapshot missed FPF by 45 minutes! I guess I'm not feeling too gracious today… | 20:01 |
| bswartz | we probably don't do a good enough job of publicizing the rules | 20:01 |
| markstur | cknight, If you go to Ben's house he may scold you first and then invite you in | 20:01 |
| *** gouthamr has quit IRC | 20:01 | |
| *** scottda has quit IRC | 20:01 | |
| bswartz | cknight: we can still FFE that if someone asks -- I was under the impression that it was tangled up with the mountable snapshot stuff which wasn't ready | 20:02 |
| markstur | There is a share/driver.py change in that patch that doesn't look good | 20:02 |
| *** rooneym has quit IRC | 20:02 | |
| *** ameade has quit IRC | 20:02 | |
| *** tpsilva has quit IRC | 20:02 | |
| *** Yogi1 has quit IRC | 20:03 | |
| cknight | bswartz: nope, not tangled up at all. But since no one else at my company seems to really care, I'm done pushing on that string. | 20:03 |
| bswartz | markstur: holy crap you're right | 20:04 |
| *** kaisers_ has quit IRC | 20:04 | |
| bswartz | when I read that code I assumed it was in the hpe driver class because it would be insane for it to be any other place | 20:04 |
| markstur | Uh-oh. I feel more censuring coming | 20:04 |
| bswartz | however he put it in the parent class | 20:04 |
| * bswartz looks for the -3 button | 20:04 | |
| bswartz | seriously though, this is easy to fix | 20:05 |
| markstur | That's -11 in binary (a Spinal Tap reference) | 20:05 |
| cknight | markstur: good catch, definitely not ready | 20:05 |
| openstackgerrit | Merged openstack/puppet-manila: Prepare 9.2.0 (newton b3) https://review.openstack.org/361209 | 20:05 |
| bswartz | markstur, cknight: those lines were not part of the change last week when I put the -2 on, they were added this week for an unknown reason | 20:06 |
| *** timcl has quit IRC | 20:06 | |
| markstur | cknight, bswartz, xyang1 Since you are here I should mention https://review.openstack.org/#/c/339795 <-- Need HPE to have update_access() in Newton. Alex addressed my last comments. | 20:12 |
| cknight | markstur: looking... | 20:13 |
| markstur | Thanks! | 20:20 |
| *** gouthamr has joined #openstack-manila | 20:23 | |
| *** akerr has joined #openstack-manila | 20:31 | |
| *** akerr_ has joined #openstack-manila | 20:35 | |
| *** akerr has quit IRC | 20:39 | |
| bswartz | anyone know where ganso went? | 20:40 |
| bswartz | he did something strange with his CI patch | 20:40 |
| openstackgerrit | Ben Swartzlander proposed openstack/manila: [CI FIX] Fix 'ip route' matching multiple subnets https://review.openstack.org/361413 | 20:43 |
| *** ameade has joined #openstack-manila | 20:43 | |
| openstackgerrit | Rodrigo Barbieri proposed openstack/manila: [CI FIX] Fix 'ip route' matching multiple subnets https://review.openstack.org/361413 | 20:45 |
| bswartz | ganso: ping | 20:45 |
| *** ganso has joined #openstack-manila | 20:46 | |
| ganso | I am back! | 20:47 |
| ganso | IRC said my account was unavailable | 20:47 |
| ganso | bswartz: we need the depends on | 20:47 |
| bswartz | ganso: that's weird | 20:47 |
| ganso | bswartz: and we need to merge mitaka patch | 20:47 |
| bswartz | but you can't depend on a stable/mitaka patch in master | 20:47 |
| ganso | bswartz: grenade runs mitaka | 20:47 |
| ganso | bswartz: it is failing in mitaka | 20:47 |
| bswartz | okay then we just merge the mitaka patch and wait | 20:48 |
| bswartz | or does the depends on magic actually fix the grenade job? | 20:48 |
| ganso | bswartz: it does fix, it worked | 20:48 |
| ganso | bswartz: I watched Zuul | 20:48 |
| bswartz | that's.... strange | 20:48 |
| ganso | bswartz: it applied the change to mitaka | 20:48 |
| ganso | bswartz: and used the change on master | 20:48 |
| ganso | bswartz: used the *updated mitaka | 20:48 |
| bswartz | okay I'll vote +2 on both of these | 20:48 |
| bswartz | anyone else want to vote or should I ninja both patches? | 20:49 |
| gouthamr | +1 ninja merge. | 20:49 |
| ganso | +1 ninja | 20:49 |
| markstur | swiftly and quietly done. Very ninja-like | 20:51 |
| bswartz | if it get's the gate moving again I'm happy to be a ninja | 20:52 |
| markstur | well if ninjas had IRC and publicly announced there intentions | 20:52 |
| markstur | +1 | 20:52 |
| ganso | bswartz: after that merges, we should recheck scenario job fix and merge that as well | 20:52 |
| bswartz | ganso: I was in the process of rebasing that | 20:53 |
| openstackgerrit | Ben Swartzlander proposed openstack/manila: Fix connectivity problem in Scenario job https://review.openstack.org/359166 | 20:53 |
| ganso | bswartz: great | 20:53 |
| *** scottda has joined #openstack-manila | 20:54 | |
| ganso | bswartz: migration patch after that one? xD | 20:56 |
| bswartz | I hope so | 20:56 |
| bswartz | the question is whether any other gate blocker bugs still lurk | 20:56 |
| ganso | bswartz: I believe generic driver still has the usual random failures, but they should be minimal after that fix merges | 20:57 |
| * bswartz sighs | 20:57 | |
| ganso | bswartz: we got to stop the merge | 21:02 |
| ganso | bswartz: forgot about the lock | 21:03 |
| ganso | bswartz: or submit another patch | 21:03 |
| *** tpsilva has joined #openstack-manila | 21:03 | |
| bswartz | why? it passed when it ran. did we just get lucky? | 21:03 |
| ganso | bswartz: lock was never there, but code will still be prone to race condition, as it always has been | 21:04 |
| bswartz | if you push another patch it will stop automatically | 21:04 |
| ganso | bswartz: the race condition is rare | 21:04 |
| ganso | bswartz: but is something to be fixed | 21:04 |
| bswartz | adding a lock can be done separately | 21:04 |
| bswartz | what symptom is exhibited if you have a race that goes badly? | 21:05 |
| ganso | bswartz: ok, I will submit another patch later tonight then | 21:05 |
| ganso | bswartz: error message will be similar to what we had | 21:05 |
| ganso | bswartz: thread lists routes, another removes, and then the first thread will fail deleting what it listed | 21:05 |
| jay-mehta | markstur, bswartz, for patch https://review.openstack.org/#/c/329552, the check in driver.py can be avoided but will require API change for all the driver in _setup_server() to accept another keyword argument. | 21:07 |
| *** eharney has quit IRC | 21:08 | |
| jay-mehta | Please suggest approach | 21:08 |
| bswartz | jay-mehta: I'm not sure what the issue is | 21:08 |
| jay-mehta | from manager.py, I am passing another keyword argument | 21:08 |
| bswartz | what is the keyword and where does it come from? | 21:08 |
| jay-mehta | It is share instance host name from manager .py which calls driver's setup_server | 21:09 |
| bswartz | you mean the backend name? | 21:09 |
| jay-mehta | which in turns forwards to driver's _setup_server() which accepts network_info, metadata=None. | 21:10 |
| jay-mehta | the kwargs in setup_server() at driver.py would pass new keyword argument "share_instance_host" which will break other driver since the API of _setup_server() does not accept arguments as *args,**kwargs | 21:11 |
| jay-mehta | We require knowledge from scheduler about host where new server is to be created which we get from this third argument | 21:12 |
| *** rooneym has joined #openstack-manila | 21:13 | |
| bswartz | jay-mehta: generally speaking, drivers need more information from the manager, the correct fix is to modify the driver interface and pass the additional information to ALL drivers | 21:13 |
| bswartz | however I'm unconvinced that this is necessary | 21:13 |
| bswartz | *if* drivers need more information from the manager | 21:14 |
| jay-mehta | Ok. I can modify API of _setup_server() to accept third argument | 21:14 |
| jay-mehta | New API would like def _setup_server(self, network_info, metadata=None, share_instance_host=None) | 21:15 |
| gouthamr | jay-mehta: where are you using that? | 21:16 |
| openstackgerrit | Rodrigo Barbieri proposed openstack/manila: Fix race condition updating routes https://review.openstack.org/361473 | 21:16 |
| bswartz | that sounds better... I dislike the use of "metadata" here | 21:16 |
| ganso | bswartz: ^ I marked as TriviialFix | 21:16 |
| gouthamr | jay-mehta: afaics you're not using the host in your driver.. | 21:16 |
| jay-mehta | gouthamr I am using it to create share server | 21:16 |
| gouthamr | jay-mehta: in any case, the driver always knows the host... self.host | 21:16 |
| bswartz | I'm still curious about what the "share_instance_host" is supposed to be and why it's needed | 21:16 |
| * gouthamr waves at ganso | 21:17 | |
| gouthamr | ganso: saw your text now. sorry was afk myself :) | 21:17 |
| gouthamr | jay-mehta: in the patch you uploaded? | 21:18 |
| bswartz | ganso: just so you know trivialfixes can't be backported -- we need a bug number | 21:18 |
| ganso | gouthamr: np :) | 21:18 |
| ganso | bswartz: alright, will update | 21:18 |
| jay-mehta | I will check one more time is I can extract host information within _setup_server() using self.host | 21:19 |
| gouthamr | jay-mehta: ah, yes | 21:19 |
| gouthamr | jay-mehta: see that now.. yes, you shouldn't need share_instance_host.. | 21:19 |
| jay-mehta | Oh great | 21:20 |
| bswartz | gouthamr: +1 | 21:20 |
| jay-mehta | I didn't knew if self.host would have host info in format: host@backend#pool inside setup_server() | 21:20 |
| jay-mehta | that simplifies a lot | 21:20 |
| bswartz | jay-mehta: you're interested in the pool name in particular? | 21:21 |
| bswartz | the share-servers feature was designed in such a way that a share server can span pools | 21:21 |
| jay-mehta | Entire host details | 21:21 |
| jay-mehta | I use pool name in specific to decide where I have to create new share server | 21:21 |
| bswartz | so tying a share-server to a pool wouldn't be correct, unless backend has a particular limitation | 21:21 |
| bswartz | jay-mehta: gouthamr is saying you have the host name already, but I don't think you have the pool name -- that's a detail of the new share being created | 21:22 |
| jay-mehta | With 3PAR, pool can have only one share server | 21:22 |
| bswartz | someone needs to looks at the share-server/pool interactions because I suspect we have bugs there | 21:23 |
| jay-mehta | hmm. I had used all the arguments of _setup_server() to see if I can extract this host information. But I didn't find one | 21:23 |
| jay-mehta | I was not aware of self.host. | 21:24 |
| gouthamr | jay-mehta bswartz: yes, share servers should span entire backends, shouldn't they.. | 21:26 |
| jay-mehta | create share server is called for each pool which does not have any share server created so far. If it has, manager utilizes it directly | 21:28 |
| jay-mehta | And this creation of share server is done using setup_server() | 21:28 |
| jay-mehta | This is where the checks are done for finding compatible_share_server | 21:30 |
| jay-mehta | in manager.py | 21:30 |
| openstackgerrit | Rodrigo Barbieri proposed openstack/manila: Fix race condition updating routes https://review.openstack.org/361473 | 21:30 |
| bswartz | gouthamr: it's meant to be up to the driver | 21:30 |
| bswartz | there's code remaining to be written there | 21:31 |
| *** gouthamr_ has joined #openstack-manila | 21:33 | |
| gouthamr_ | bswartz jay-mehta: right.. so share servers belong to host@backend - not pools, atleast not so far.. | 21:34 |
| gouthamr_ | jay-mehta: so, if you make the change to create a share server for a pool, manila's share manager may still use something that belongs to another pool as long as there is one.. | 21:35 |
| *** gouthamr has quit IRC | 21:36 | |
| jay-mehta | gouthamr_ , bswartz says it is specific to the driver | 21:36 |
| openstackgerrit | Rodrigo Barbieri proposed openstack/manila: Fix race condition updating routes https://review.openstack.org/361473 | 21:37 |
| gouthamr_ | jay-mehta: look at "get_available_share_servers" | 21:37 |
| jay-mehta | checking | 21:37 |
| ganso | bswartz: sorry for so many edits. Doing everything from my phone | 21:38 |
| gouthamr_ | jay-mehta: you'd have to override "choose_share_server_compatible_with_share" and disqualify any share servers that don't belong to a pool in your case. | 21:39 |
| gouthamr_ | jay-mehta: that don't belong to the requested pool.. | 21:39 |
| jay-mehta | gouthamr_, I think so yes. | 21:40 |
| gouthamr_ | jay-mehta: hmmm, the sanest thing to do is put it inside the metadata field so that you can avoid all the drivers breaking (bswartz doesn't seem to agree though) | 21:44 |
| jay-mehta | gouthamr_ in get_available_share_server, we make a query in db passing self.host. If the host will have pool information in form: host@backend#pool, will it not be specific to a pool? | 21:45 |
| gouthamr_ | jay-mehta: nope.. it's just host@backend | 21:46 |
| gouthamr_ | you'll get the pool info that you need in 'choose_share_server_compatible_with_share' with the list of share servers that your backend has | 21:46 |
| gouthamr_ | jay-mehta: so you'll have to find the best one yourself.. | 21:47 |
| *** kaisers_ has joined #openstack-manila | 21:48 | |
| *** dustins_ has quit IRC | 21:48 | |
| jay-mehta | gouthamr_, So are you suggesting that I should override choose_share_server_compatible_with_share() and use share_instance to extract pool information and use it accordingly for setup_server? | 21:49 |
| *** akerr_ has quit IRC | 21:51 | |
| gouthamr_ | jay-mehta: i guess this can work... you can add the share_instance's host in the metadata in _setup_server, so far, it's not being used.. don't call it share_instance_host though, prefer just 'request_host' or something like that.. | 21:52 |
| *** kaisers_ has quit IRC | 21:52 | |
| gouthamr_ | jay-mehta: and override "choose_share_server_compatible_with_share" to look at the list of share servers and find the one that corresponds to the requested share['host'] | 21:53 |
| *** alyson_ has quit IRC | 21:53 | |
| gouthamr_ | jay-mehta: share['host'] will be a fully qualified host string host@backend#pool | 21:53 |
| *** gouthamr_ is now known as gouthamr | 21:57 | |
| *** xyang1 has quit IRC | 21:57 | |
| jay-mehta | gouthamr, ok. And should I pass information as dictionary in metadata? | 21:58 |
| gouthamr | jay-mehta: yes, it was meant to be a dict.. | 21:58 |
| *** JoseMello has quit IRC | 22:00 | |
| jay-mehta | gouthamr, Thanks . I will use it as metadata['request_host']. Yes that will not break any other driver and won't require introducing any other argument. | 22:00 |
| gouthamr | jay-mehta: you're welcome.. | 22:01 |
| jay-mehta | gouthamr, I still see that in manager.py, metadata is not passed to _setup_server(). https://github.com/openstack/manila/blob/master/manila/share/manager.py | 22:03 |
| jay-mehta | Ling 468 | 22:04 |
| openstackgerrit | Merged openstack/manila: [CI FIX] Fix 'ip route' matching multiple subnets https://review.openstack.org/361413 | 22:04 |
| gouthamr | jay-mehta: yes, you can pass it, it's a kwarg in ShareDriver | 22:04 |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/manila: Updated from global requirements https://review.openstack.org/360891 | 22:07 |
| jay-mehta | gouthamr, I am pointing out metadata argument in _setup_server(), Line 468 in manager.py . I see that caller does not pass metadata. Am I missing | 22:10 |
| gouthamr | jay-mehta: not sure what the question is.. i meant you'll have to make the change to fill it in with request_host and pass it.. | 22:11 |
| jay-mehta | The manager api accepts metadata which can be passed on to driver's setup_server. But the caller at line 467 in manager.py does not pass metadata | 22:13 |
| *** esker has joined #openstack-manila | 22:15 | |
| jay-mehta | gouthamr, If I am not wrong, the flow is from line 467 to line 2344 in manager.py . But here metadata is not passed | 22:16 |
| *** esker has quit IRC | 22:18 | |
| gouthamr | jay-mehta: https://review.openstack.org/#/c/329552/9/manila/share/manager.py | 22:20 |
| gouthamr | jay-mehta: create the metadata dict with request_host in it and pass it on L#467 ... | 22:21 |
| jay-mehta | Yes | 22:21 |
| jay-mehta | I will have to begin with creating a new one | 22:22 |
| jay-mehta | New metadata dict at line 467. Thats what I was saying. | 22:22 |
| jay-mehta | Got it. Thanks gouthamr | 22:22 |
| jay-mehta | Thanks bswartz. | 22:23 |
| jay-mehta | gouthamr, I had updated unit test for some share and generic driver and Huawei . https://review.openstack.org/#/c/360088/ If you could please review this | 22:27 |
| *** rooneym has quit IRC | 23:02 | |
| *** scottda has quit IRC | 23:02 | |
| *** cknight has quit IRC | 23:02 | |
| *** ameade has quit IRC | 23:03 | |
| *** tpsilva has quit IRC | 23:03 | |
| *** lpetrut has quit IRC | 23:08 | |
| *** DericHorn-HP has quit IRC | 23:17 | |
| *** rooneym has joined #openstack-manila | 23:18 | |
| *** DericHorn-HP has joined #openstack-manila | 23:25 | |
| *** kaisers_ has joined #openstack-manila | 23:37 | |
| *** amit213 has quit IRC | 23:37 | |
| *** amit213 has joined #openstack-manila | 23:40 | |
| *** kaisers_ has quit IRC | 23:41 | |
| *** rooneym has quit IRC | 23:46 | |
| *** ganso has quit IRC | 23:46 | |
| *** DericHorn-HP has quit IRC | 23:58 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!