| *** yangyapeng has quit IRC | 00:05 | |
| *** Suyi has quit IRC | 00:21 | |
| *** Suyi has joined #openstack-manila | 00:42 | |
| *** gouthamr has joined #openstack-manila | 00:52 | |
| *** zengyingzhe_ has quit IRC | 00:56 | |
| *** zengyingzhe has joined #openstack-manila | 00:59 | |
| *** yangyapeng has joined #openstack-manila | 01:21 | |
| *** wangqun has joined #openstack-manila | 01:31 | |
| *** yangyape_ has joined #openstack-manila | 01:32 | |
| *** yangyapeng has quit IRC | 01:35 | |
| *** yangyape_ has quit IRC | 01:49 | |
| *** vbellur has joined #openstack-manila | 02:00 | |
| *** mtanin___ has quit IRC | 02:07 | |
| *** yangyapeng has joined #openstack-manila | 02:36 | |
| *** yangyape_ has joined #openstack-manila | 02:37 | |
| *** yangyapeng has quit IRC | 02:40 | |
| *** yangyape_ has quit IRC | 02:47 | |
| *** yangyape_ has joined #openstack-manila | 02:50 | |
| *** yangyap__ has joined #openstack-manila | 02:56 | |
| *** gouthamr_ has joined #openstack-manila | 02:57 | |
| *** gouthamr has quit IRC | 02:58 | |
| *** yangyape_ has quit IRC | 02:58 | |
| *** yangyapeng has joined #openstack-manila | 03:01 | |
| *** gouthamr_ is now known as gouthamr | 03:01 | |
| *** yangyap__ has quit IRC | 03:05 | |
| *** yangyapeng has quit IRC | 03:17 | |
| *** yangyapeng has joined #openstack-manila | 03:22 | |
| *** Suyi has quit IRC | 03:25 | |
| *** yangyape_ has joined #openstack-manila | 03:34 | |
| *** yangyapeng has quit IRC | 03:36 | |
| *** ganso has quit IRC | 03:41 | |
| *** gouthamr has quit IRC | 03:44 | |
| *** daidv has joined #openstack-manila | 03:49 | |
| *** yangyape_ has quit IRC | 03:50 | |
| *** daidv has quit IRC | 04:07 | |
| *** Suyi has joined #openstack-manila | 04:23 | |
| *** assassin has quit IRC | 04:32 | |
| *** yangyape_ has joined #openstack-manila | 04:37 | |
| openstackgerrit | Ramana Raja proposed openstack/python-manilaclient: add access_key to access_list API's response https://review.openstack.org/344948 | 04:39 |
|---|---|---|
| *** yangyape_ has quit IRC | 04:45 | |
| *** yangyape_ has joined #openstack-manila | 04:50 | |
| *** yangyape_ has quit IRC | 04:55 | |
| *** yangyape_ has joined #openstack-manila | 05:04 | |
| *** yangyape_ has quit IRC | 05:09 | |
| *** Suyi has quit IRC | 05:12 | |
| *** yangyape_ has joined #openstack-manila | 05:20 | |
| *** yangyape_ has quit IRC | 05:28 | |
| *** yangyape_ has joined #openstack-manila | 05:33 | |
| *** sandanar has joined #openstack-manila | 05:37 | |
| *** nkrinner_afk is now known as nkrinner | 05:48 | |
| *** yangyape_ has quit IRC | 05:50 | |
| *** yangyape_ has joined #openstack-manila | 05:59 | |
| *** yangyap__ has joined #openstack-manila | 06:00 | |
| *** yangyap__ has quit IRC | 06:01 | |
| *** yangyap__ has joined #openstack-manila | 06:02 | |
| *** yangyapeng has joined #openstack-manila | 06:03 | |
| openstackgerrit | Alexey Ovchinnikov proposed openstack/manila: [DNM] Tests for scenario tests https://review.openstack.org/340236 | 06:04 |
| *** yangyape_ has quit IRC | 06:05 | |
| *** bhakta_ has quit IRC | 06:05 | |
| *** yangyapeng has quit IRC | 06:07 | |
| *** yangyap__ has quit IRC | 06:07 | |
| *** bhakta has joined #openstack-manila | 06:12 | |
| *** pcaruana has joined #openstack-manila | 06:14 | |
| openstackgerrit | Thomas Bechtold proposed openstack/manila: Remove enable_v1_api and enable_v2_api config opts https://review.openstack.org/358406 | 06:27 |
| *** lpetrut has joined #openstack-manila | 06:35 | |
| *** yangyape_ has joined #openstack-manila | 06:38 | |
| *** ubijtsa has joined #openstack-manila | 06:40 | |
| *** ubijtsa is now known as Guest11182 | 06:40 | |
| *** Guest11182 has quit IRC | 06:41 | |
| *** Guest11182 has joined #openstack-manila | 06:41 | |
| *** yangyapeng has joined #openstack-manila | 06:42 | |
| *** yangyape_ has quit IRC | 06:42 | |
| *** Guest11182 is now known as assassin | 06:42 | |
| *** yangyape_ has joined #openstack-manila | 06:44 | |
| *** yangyapeng has quit IRC | 06:48 | |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/manila-ui: Imported Translations from Zanata https://review.openstack.org/358985 | 06:51 |
| *** yangyape_ has quit IRC | 07:01 | |
| openstackgerrit | NidhiMittalHada proposed openstack/manila: manila does not calculate allocated_capacity_gb https://review.openstack.org/351034 | 07:03 |
| *** dsariel has joined #openstack-manila | 07:04 | |
| *** yangyapeng has joined #openstack-manila | 07:05 | |
| *** qeelee has joined #openstack-manila | 07:21 | |
| *** lpetrut has quit IRC | 07:26 | |
| *** lpetrut has joined #openstack-manila | 07:40 | |
| *** lpetrut has quit IRC | 08:01 | |
| *** openstackgerrit has quit IRC | 08:03 | |
| *** openstackgerrit has joined #openstack-manila | 08:04 | |
| *** aovchinnikov has joined #openstack-manila | 08:17 | |
| *** dsariel has quit IRC | 08:20 | |
| *** dsariel has joined #openstack-manila | 08:26 | |
| *** qeelee has quit IRC | 08:30 | |
| *** lpetrut has joined #openstack-manila | 08:41 | |
| *** lpetrut has quit IRC | 08:50 | |
| *** kaisers_ has joined #openstack-manila | 09:10 | |
| *** lpetrut has joined #openstack-manila | 09:21 | |
| openstackgerrit | zhongjun proposed openstack/manila: Implement share backup https://review.openstack.org/343980 | 09:28 |
| *** nkrinner has quit IRC | 09:28 | |
| *** nkrinner has joined #openstack-manila | 09:42 | |
| *** wangqun has quit IRC | 10:03 | |
| *** qeelee has joined #openstack-manila | 10:05 | |
| *** lpetrut1 has joined #openstack-manila | 10:10 | |
| *** lpetrut has quit IRC | 10:11 | |
| *** lpetrut1 is now known as lpetrut | 10:11 | |
| *** yangyapeng has quit IRC | 10:28 | |
| openstackgerrit | Cedric Zhuang proposed openstack/manila: Add documentation for EMC Unity Driver for Manila https://review.openstack.org/358576 | 10:33 |
| *** qeelee has quit IRC | 10:43 | |
| openstackgerrit | Rodrigo Barbieri proposed openstack/manila-ui: Add share type change to Share Migration UI https://review.openstack.org/345664 | 10:50 |
| openstackgerrit | Rodrigo Barbieri proposed openstack/manila-ui: Add Share Migration support https://review.openstack.org/344431 | 10:50 |
| *** ganso has joined #openstack-manila | 10:50 | |
| *** alyson_ has joined #openstack-manila | 11:06 | |
| *** JoseMello has joined #openstack-manila | 11:21 | |
| openstackgerrit | Alexey Khodos proposed openstack/manila: Nexenta: adding share drivers for NexentaStor https://review.openstack.org/309286 | 11:30 |
| openstackgerrit | Tatiana Ovchinnikova proposed openstack/manila-ui: Additional fields for Create/Update Share Type forms https://review.openstack.org/338504 | 11:33 |
| openstackgerrit | Merged openstack/manila-ui: Imported Translations from Zanata https://review.openstack.org/358985 | 11:33 |
| *** netapp has quit IRC | 11:41 | |
| *** netapp has joined #openstack-manila | 11:48 | |
| *** wangqun has joined #openstack-manila | 11:51 | |
| *** dsariel has quit IRC | 11:58 | |
| *** yangyapeng has joined #openstack-manila | 12:01 | |
| *** xyang1 has joined #openstack-manila | 12:05 | |
| *** tpsilva has joined #openstack-manila | 12:08 | |
| *** kaisers_ has quit IRC | 12:11 | |
| zhongjun_ | bswartz: ping | 12:14 |
| openstackgerrit | Rodrigo Barbieri proposed openstack/manila: [DEBUG] PLEASE DO NOT MERGE 4 https://review.openstack.org/359166 | 12:26 |
| *** gouthamr has joined #openstack-manila | 12:26 | |
| tpsilva | gouthamr: ping | 12:41 |
| gouthamr | hey tpsilva | 12:41 |
| tpsilva | gouthamr: hey Goutham! good morning | 12:41 |
| tpsilva | gouthamr: quick question: how can I find the bug issues raised by the DocImpact flag? | 12:41 |
| gouthamr | tpsilva: hey.. they will have a DocImpact right on top.. | 12:51 |
| tpsilva | gouthamr: oh, but the DocImpact flag does not raise an issue on launchpad or something? | 12:51 |
| gouthamr | tpsilva: https://bugs.launchpad.net/manila?field.searchtext=DocImpact&search=Search&field.status%3Alist=NEW&field.status%3Alist=INCOMPLETE_WITH_RESPONSE&field.status%3Alist=INCOMPLETE_WITHOUT_RESPONSE&field.status%3Alist=CONFIRMED&field.status%3Alist=TRIAGED&field.status%3Alist=INPROGRESS&field.status%3Alist=FIXCOMMITTED&field.assignee=&field.bug_reporter=&field.omit_dupes=on&field.has_patch=&field.has_no_package= | 12:51 |
| tpsilva | gouthamr: ah cool, it's right on manila's LP | 12:52 |
| tpsilva | gouthamr: thanks | 12:52 |
| gouthamr | tpsilva: you're welcome.. | 12:52 |
| *** chlong has quit IRC | 12:56 | |
| ganso | bswartz: ping | 12:57 |
| *** dustins has joined #openstack-manila | 13:00 | |
| *** porrua has joined #openstack-manila | 13:11 | |
| *** chlong has joined #openstack-manila | 13:12 | |
| *** qeelee has joined #openstack-manila | 13:13 | |
| *** wangqun has quit IRC | 13:14 | |
| *** qeelee has quit IRC | 13:15 | |
| *** akerr has joined #openstack-manila | 13:21 | |
| *** timcl has joined #openstack-manila | 13:24 | |
| *** rraja has joined #openstack-manila | 13:26 | |
| *** nkrinner is now known as nkrinner_afk | 13:35 | |
| *** sandanar has quit IRC | 13:45 | |
| bswartz | zhongjun_, ganso: pong | 13:46 |
| bswartz | sorry I was in a meeting this morning | 13:46 |
| bswartz | I'm out now | 13:46 |
| ganso | bswartz: Hi Ben | 13:46 |
| ganso | bswartz: I read your comment | 13:46 |
| bswartz | https://review.openstack.org/#/c/332267/56 | 13:47 |
| ganso | bswartz: are you suggesting adding a --disruptable API parameter? | 13:47 |
| bswartz | ganso: well the driver interface could change later so it's not urgent to fix that | 13:47 |
| bswartz | It just occurred to me that we don't need a bunch of different methods to check various migration capabitilies | 13:47 |
| ganso | bswartz: indeed | 13:47 |
| bswartz | because you can pass all the requested capabilities down to 1 method and get back a yes or no answer | 13:48 |
| ganso | bswartz: I was talking to goutham | 13:48 |
| ganso | bswartz: we can combine them all in migration_check_compatibility | 13:48 |
| ganso | bswartz: but you suggested something useful to signal the manager that the migration is going to be disruptive or not | 13:48 |
| ganso | bswartz: goutham told me he can check that | 13:48 |
| bswartz | well it's up to the user | 13:49 |
| gouthamr | administrator* | 13:49 |
| bswartz | if the user requests a nondisruptive migration, we should either give him a nondisruptive migration or fail | 13:49 |
| bswartz | if the user doesn't request one, then he shouldn't get one | 13:49 |
| ganso | bswartz: but we have no use for that information currently. We previously had 1-phase/2-phase migration with the purpose of controlling that, while we always expect migration to be disruptive | 13:49 |
| ganso | bswartz: yes, exactly how writable and preserve-metadata works | 13:49 |
| ganso | bswartz: I just wanted to confirm with you because we raised this topic in midcycle, and we previously agreed we did not really feel the need of that parameter | 13:50 |
| ganso | bswartz: so looks like now we changed our minds and we want the admin to say "I want a non-disruptive"... if it fails saying "driver cannot do non-disruptive" (and fallback obviously cannot as well), then admin can retry with "ok I want a disruptive" | 13:51 |
| bswartz | ganso: do you remember why? | 13:51 |
| bswartz | I wonder if we're confusing the 2-phase vs. 1-phase migration debate with the non-disruptive question | 13:52 |
| ganso | bswartz: I think it was because we always expected migration to be disruptive, while share-modify (aka retype) could be non-disruptive if it does not invoke migration | 13:52 |
| bswartz | or the issue of 2 share instances vs 1 share instance for driver-assisted migration | 13:52 |
| ganso | bswartz: I believe we are not confusing with these other topics at this moment | 13:52 |
| bswartz | I feel strongly that we should always have 2 phase migration and 2 share instances no matter what happens internally | 13:52 |
| ganso | bswartz: I changed the code to reflect that | 13:53 |
| bswartz | however if the migration is nondisruptive then the second phase might be a no-op | 13:53 |
| ganso | bswartz: if it is non-disruptive, we could have 1-phase... | 13:53 |
| *** eharney has joined #openstack-manila | 13:53 | |
| ganso | bswartz: but we removed that possibility | 13:53 |
| openstackgerrit | Alexey Ovchinnikov proposed openstack/manila: [Do Not Merge] Generic driver ports test https://review.openstack.org/359209 | 13:53 |
| bswartz | right, we could, but for API consistency we shouldn't | 13:53 |
| ganso | aovchinnikov: #link https://review.openstack.org/359166 <== possible fix, it is zuul right now | 13:54 |
| ganso | aovchinnikov: it is *in zuul right now | 13:54 |
| ganso | bswartz: yes | 13:55 |
| ganso | bswartz: so, do we agree that we want a '--disruptive' parameter? | 13:55 |
| bswartz | it should be --non-disruptive, defaulting to false | 13:56 |
| bswartz | or maybe --nondisruptive | 13:56 |
| ganso | bswartz: ok so the default is to be disruptive, and if admin strongly needs share to be migrated non-disruptively, then he specifies it explicitely | 13:57 |
| ganso | bswartz: correct? | 13:57 |
| bswartz | yeah but it applies to other migration-like APIs just the same | 13:59 |
| bswartz | an end user could request a nondisruptive retype | 13:59 |
| *** kaisers_ has joined #openstack-manila | 13:59 | |
| ganso | bswartz: yes, we had already agreed to that when we discussed share-retype/modify API. The default was going to be disruptive, according to that etherpad | 14:02 |
| *** pcaruana has quit IRC | 14:02 | |
| zhongjun_ | bswartz: ping, In https://review.openstack.org/#/c/312321/ | 14:03 |
| bswartz | I'm just pointing out that all of the options for migration should be common across all migration-like APIs | 14:03 |
| zhongjun_ | bswartz: tbarron think we could use ipaddress.IPv6Network() and IPv4Network instead of our validation function. It means whether to allow host bits when specifying which networks can access a share | 14:03 |
| zhongjun_ | eg: 10.0.0.2/16 is invalid when we use IPv4Network. 10.0.0.2/16 is valid when we use our validation function | 14:03 |
| bswartz | zhongjun_: ah I'm glad you caught me before you went to sleep | 14:03 |
| zhongjun_ | bswartz: :D | 14:03 |
| *** kaisers_ has quit IRC | 14:04 | |
| ganso | bswartz: ok, I will make the changes. The disruptive API flag one will be newton-migration-improvements flag because it changes API, not the one you -1'ed | 14:04 |
| bswartz | ganso: my -1 was about the LVM change not that suggestion | 14:04 |
| *** tbarron|gone is now known as tbarron | 14:05 | |
| bswartz | zhongjun_: why is that string invalid for IPv4Network? it looks valid to me | 14:05 |
| tbarron | bswartz: it is ambiguous | 14:06 |
| bswartz | oh I see | 14:06 |
| bswartz | it should have been 10.0.0.0/16 | 14:06 |
| bswartz | the .2 is the problem | 14:06 |
| zhongjun_ | bswartz: yes | 14:06 |
| tbarron | 10.0.0.2/16 as an *access rule*, not as a host address and network spec | 14:06 |
| bswartz | yes | 14:06 |
| bswartz | well then we have a bug and we can fix it | 14:06 |
| tbarron | does it mean 10.0.0.0/16 or 10.0.0.0/32? | 14:06 |
| tbarron | bswartz: +1 | 14:07 |
| gouthamr | tbarron: it should not... | 14:07 |
| bswartz | tbarron: there's no way it means the second one | 14:07 |
| bswartz | it has to mean the first one, but it's incorrectly specified | 14:07 |
| tbarron | bswartz: not to you | 14:07 |
| gouthamr | bswartz: tested zhongjun_'s cases last night.. we fail at the driver.. | 14:07 |
| gouthamr | bswartz: and if you added a bunch of rules, it makes reconciling very difficult.. | 14:08 |
| bswartz | zhongjun_: I think we should use IPv4Network, it sounds like it does the right thing | 14:08 |
| gouthamr | so +1 for API check | 14:08 |
| tbarron | in any case there should be one way to specify an access rule and it should cover the number of bits left to right with no host bits | 14:08 |
| tbarron | use /32 if you want exactly one host | 14:08 |
| gouthamr | using /32 is still two hosts | 14:10 |
| tbarron | otherwise you can have multiple rules like 10.0.0.0/16, 10.0.0.2/16, 10.0.5.5/16 in one access list and they all mean the same thing (if that's the way we "have" to interpret it) | 14:10 |
| zhongjun_ | bswartz: Is there have any reason we use the ourselves code to check the ip format? | 14:10 |
| tbarron | the ipaddress module didn't used to exist in python2 | 14:10 |
| bswartz | tbarron: that's a problem | 14:11 |
| bswartz | tbarron: we can't drop support for py2.7 yet | 14:11 |
| tbarron | it does exist in python2.7 | 14:11 |
| tbarron | not in 2.6 | 14:11 |
| bswartz | o_O | 14:11 |
| tbarron | that's why we rolled our own code i hypothesize | 14:11 |
| bswartz | tbarron: possible, but that sucks | 14:12 |
| *** alyson_ has quit IRC | 14:13 | |
| tbarron | gouthamr: /30 is two hosts, 4 addresses /31 2 addresses, /32 1 adedress | 14:13 |
| tbarron | anyways IMO we shouldn't roll our own IP validation code,.it's done in standard ways by a much larger community | 14:14 |
| tbarron | the chances of doing ipv6 wrong are too high | 14:14 |
| *** cknight has joined #openstack-manila | 14:15 | |
| tbarron | there is a standard module and we should use it | 14:15 |
| *** pcaruana has joined #openstack-manila | 14:17 | |
| openstackgerrit | Alyson proposed openstack/manila: Validate EVS IP in Manila HNAS Driver https://review.openstack.org/359222 | 14:17 |
| zhongjun_ | But how to deal with the python version. Is it a problem? | 14:18 |
| bswartz | tbarron: /31 is invalid | 14:18 |
| bswartz | only 0-30 and 32 are valid | 14:19 |
| bswartz | (for ipv4) | 14:19 |
| zhongjun_ | Maybe we could save netutils.is_valid_ipv6 check in N version, change to IPv6Network and IPv4Nework in next O version. | 14:19 |
| bswartz | honestly /30 is debatable too, /29 is the smallest that most people go | 14:19 |
| *** aovchinnikov has quit IRC | 14:21 | |
| gouthamr | is python2.6 even supported? we run py27 and py34 | 14:22 |
| *** dsariel has joined #openstack-manila | 14:23 | |
| gouthamr | Support Python 2.6 was dropped in Juno | 14:24 |
| gouthamr | https://wiki.openstack.org/wiki/Python3#Python_2:_Python_2.6_support_dropped.2C_Python_2.7_only | 14:24 |
| zhongjun_ | I didn't see python2.6 in Jenkins | 14:24 |
| gouthamr | so, we can bravely use the module tbarron proposed ? | 14:25 |
| tbarron | bswartz: /31 is not invalid as access list!! | 14:25 |
| tbarron | bswartz: it's not usable as a network with broadcast and gateway though | 14:25 |
| rraja | tbarron ganso gouthamr : can you look at https://review.openstack.org/#/c/343306/ and https://review.openstack.org/#/c/344948 ? the CIs for both manila and the dependent manila-client patch are passing now. | 14:28 |
| gouthamr | zhongjun_: to answer your question, i strongly think you can use the validation tbarron proposed in your ipv6 patch.. bswartz -2ed the patch from tbarron though.. | 14:28 |
| gouthamr | rraja | 14:29 |
| gouthamr | sure | 14:29 |
| *** sage___ has quit IRC | 14:30 | |
| tbarron | my patch was just for reference purposes, and perhps for revival in O | 14:30 |
| rraja | gouthamr: :) | 14:30 |
| cknight | tbarron: I completely agree we shouldn't be using any of our own IP address validation code. | 14:31 |
| zhongjun_ | so, everyone agree with we need to use IPv4Network and IPv6Network function to check the ip format? | 14:31 |
| *** timcl has quit IRC | 14:31 | |
| cknight | tbarron: I would also ask whether we need a separate ipv6 access type, or whether the address type can be inferred from the value. | 14:32 |
| *** sage has joined #openstack-manila | 14:32 | |
| bswartz | zhongjun_: yes | 14:32 |
| tbarron | rraja: I see that it's all passing now - nice! Will look whan I get out of a meeting in an hour or so | 14:32 |
| zhongjun_ | cool, I will do it. | 14:32 |
| rraja | tbarron: sounds great. | 14:32 |
| tbarron | cknight: the module also has a ip_address method that works for both ipv4 and ipv6 | 14:32 |
| cknight | tbarron: Yes, so why do we need an ipv6 access type? | 14:33 |
| bswartz | we should be careful about IPv6 | 14:33 |
| bswartz | it has NO test coverage and if we start supporting it it could introduce all kind of bugs | 14:34 |
| bswartz | I would want the test coverage first before we support ipv6 | 14:34 |
| tbarron | cknight: I agree but what you are suggesing was a more radical change :) | 14:34 |
| cknight | bswartz: Agreed, ipv6 isn't to be taken lightly. | 14:34 |
| cknight | tbarron: Not radical, imho. :-) | 14:34 |
| tbarron | cknight: not radical, but more churn against the review for ipv6 that has been proposed | 14:36 |
| cknight | tbarron: Well, if the the ipv6 type isn't needed, then the review should stop. Done. | 14:37 |
| cknight | tbarron: I guess that is radical :-) | 14:37 |
| gouthamr | zhongjun_: have you tested this against any gate driver? https://review.openstack.org/#/c/312321/23/manila_tempest_tests/tests/api/test_rules.py | 14:47 |
| zhongjun_ | gouthamr: I tested it in huawei driver | 14:48 |
| gouthamr | zhongjun_: hmmm, i can try testing it against the NetApp driver. but i think having one of the gate drivers running the tests would be beneficial. | 14:49 |
| cknight | gouthamr: s/beneficial/essential/ | 14:49 |
| zhongjun_ | gouthamr: I already run it in another patch | 14:49 |
| gouthamr | cknight: +1 | 14:49 |
| zhongjun_ | https://review.openstack.org/357320 | 14:50 |
| gouthamr | i guess the argument for allowing ipv6 as another type of access rule is to allow drivers to make any changes if necessary | 14:50 |
| gouthamr | though if all backends can support it, universally, then it's just another kind of access rule | 14:51 |
| gouthamr | ip access rule | 14:51 |
| zhongjun_ | cknight: Hi, I think the ipv4 and ipv6 are different format. If we add ipv6 access type, it will make user more clear | 14:51 |
| cknight | zhongjun_: Yes, they are a different format. So the code can figure it out for itself. | 14:52 |
| *** Yogi1 has joined #openstack-manila | 14:53 | |
| zhongjun_ | cknight: and make the driver easy to distinguish between the ipv4 and ipv6 | 14:53 |
| cknight | zhongjun_: We're here to make user's lives easier, not driver developers. | 14:53 |
| gouthamr | zhongjun_: fair point.. but cknight is suggesting making the change internally in manila | 14:53 |
| gouthamr | zhongjun_: but still maintain that it is an ip access rule.. rather than an ipv6 in the API.. | 14:54 |
| *** timcl has joined #openstack-manila | 14:55 | |
| cknight | gouthamr, zhongjun_: I would be OK with ipv6 as a driver capability, since not all will support it initially, so the API could reject ipv6 rules for some shares. But I don't think a new access type is warranted here. | 14:56 |
| openstackgerrit | Goutham Pacha Ravi proposed openstack/manila: [DNM] Test IPv6 rules with all drivers https://review.openstack.org/359268 | 15:02 |
| *** vbellur has quit IRC | 15:02 | |
| gouthamr | zhongjun_: i see that your driver doesn't need any changes | 15:02 |
| gouthamr | zhongjun_:lets see how other drivers fare: https://review.openstack.org/359268 | 15:02 |
| zhongjun_ | cknight: there are many capabilities in manila, It will be little wiered, We only have ipv6 capability, the other share types not needed(such as cert, ceph) | 15:02 |
| zhongjun_ | s/share types/access_types | 15:03 |
| lpetrut | Hey guys, I've sent a chain of patches on the Windows SMB driver, fixing some access right related issues, as well as implementing the 'update_access' method. I'd be grateful if somebody could take a look: https://review.openstack.org/#/c/354855 https://review.openstack.org/#/c/354854 https://review.openstack.org/#/c/354853 | 15:03 |
| cknight | zhongjun_: Those are for specific protocols where existing access types wouldn't work. Someone needs to explain to me why 'ip' isn't good enough for both ipv4 and ipv6 addresses when existing libraries can easily handle both and distinguish between them. | 15:05 |
| zhongjun_ | gouthamr: ok | 15:07 |
| gouthamr | zhongjun_: some drivers might fail on CIFS ipv6 tests.. which, i don't think make sense.. though i'm not an expert :) | 15:08 |
| bswartz | oh I see this discussion didn't end | 15:09 |
| zhongjun_ | cknight: Only use ip to include ipv4 and ipv6 will be more easy for me | 15:09 |
| bswartz | personally I'd like to see IPv6 support across all manila and all the drivers | 15:10 |
| bswartz | I think I agree that "ip" can cover both because it's easy to disambiguate between them | 15:10 |
| bswartz | where you get into trouble is with URL-like strings that can include port numbers | 15:10 |
| bswartz | export locations might need to take the URL-like IPv6 format with brackets | 15:11 |
| bswartz | however access rules should be easy | 15:11 |
| bswartz | the main thing is proper test coverage before we claim support (and everything passing those tests) | 15:12 |
| zhongjun_ | gouthamr: I still not use the CIFS ipv6 tests in huawei driver, it will be fail in huawei CI | 15:16 |
| zhongjun_ | gouthamr: my driver changes for ipv6 in following link: https://review.openstack.org/#/c/357320/1/manila/share/drivers/huawei/v3/connection.py | 15:26 |
| *** vbellur has joined #openstack-manila | 15:27 | |
| gouthamr | zhongjun_: :( i see | 15:28 |
| gouthamr | zhongjun_: so the rule type is ipv6.. | 15:28 |
| gouthamr | zhongjun_: many drivers check that before applying.. so, that test patch will not work on most drivers. | 15:29 |
| zhongjun_ | gouthamr: yes, If we conbine ipv4 and ipv6, my driver doesn't need any changes | 15:29 |
| zhongjun_ | combine | 15:30 |
| gouthamr | zhongjun_: yes.. | 15:30 |
| gouthamr | zhongjun_: I suggest having a test patch with the access_type='ip', without all the validation code you added, to see if the positive tests you wrote work for all the gate drivers.. | 15:32 |
| zhongjun_ | gouthamr: yes, it will be better | 15:35 |
| *** alyson_ has joined #openstack-manila | 15:47 | |
| *** kaisers_ has joined #openstack-manila | 15:48 | |
| openstackgerrit | Alyson proposed openstack/manila: Validate EVS IP in Manila HNAS Driver https://review.openstack.org/359222 | 15:50 |
| *** kaisers_ has quit IRC | 15:53 | |
| *** pcaruana has quit IRC | 15:56 | |
| openstackgerrit | Rodrigo Barbieri proposed openstack/manila: Fix connectivity problem in Scenario job https://review.openstack.org/359166 | 15:59 |
| openstackgerrit | zhongjun proposed openstack/manila: [DNM] Test IP(ipv4 and ipv6) rules with all drivers https://review.openstack.org/359316 | 16:08 |
| *** pcaruana has joined #openstack-manila | 16:09 | |
| openstackgerrit | zhongjun proposed openstack/manila: [DNM] Test IP(ipv4 and ipv6) rules with all drivers https://review.openstack.org/359316 | 16:20 |
| *** Yogi1 has quit IRC | 16:27 | |
| *** eharney has quit IRC | 16:35 | |
| *** Suyi has joined #openstack-manila | 16:36 | |
| *** rraja has quit IRC | 16:38 | |
| openstackgerrit | Rodrigo Barbieri proposed openstack/manila: Fix Share Migration improper behavior for drivers https://review.openstack.org/332267 | 16:56 |
| *** mtanino has joined #openstack-manila | 16:57 | |
| bswartz | gouthamr: ping | 16:59 |
| gouthamr | bswartz: pong | 16:59 |
| bswartz | ahm ended | 17:00 |
| ganso | gouthamr: oh yea, forgot about that keyerror | 17:00 |
| * gouthamr walks over | 17:00 | |
| ganso | gouthamr: did you return an empty dict or did you not implement the method (thus it raised NotImplementedError)? | 17:05 |
| ganso | gouthamr: I assume you returned an empty dict | 17:08 |
| *** mtanin___ has joined #openstack-manila | 17:10 | |
| *** mtanino has quit IRC | 17:12 | |
| *** dsariel has quit IRC | 17:13 | |
| gouthamr | ganso: nope | 17:16 |
| ganso | gouthamr: I was looking at the wrong code version | 17:17 |
| gouthamr | ganso: i returned a dict without those keys.. lemme grab that from my logs | 17:17 |
| ganso | gouthamr: I guess you do have 'total_progress' set | 17:22 |
| gouthamr | ganso: yes i do... | 17:23 |
| *** cfey_ has quit IRC | 17:23 | |
| *** cfey has joined #openstack-manila | 17:24 | |
| *** dsariel has joined #openstack-manila | 17:27 | |
| gouthamr | ganso: also, when a migration is not in progress, i get this message | 17:30 |
| gouthamr | $ manila migration-complete test | 17:30 |
| gouthamr | ERROR: Invalid share: First migration phase of share 7228b35b-a541-4805-8c27-ced5f231d5d7 not completed yet. | 17:30 |
| ganso | gouthamr: did you return True in migration_continue? | 17:31 |
| *** rraja has joined #openstack-manila | 17:32 | |
| gouthamr | ganso: yeah; the task state was 'migration_success' | 17:32 |
| gouthamr | i didn't specify 'notify'... | 17:33 |
| *** lpetrut has quit IRC | 17:33 | |
| gouthamr | ganso: which will go away.. but either case, when the task state is something else, guess we should use an appropriate error message | 17:33 |
| ganso | gouthamr: notify functionality has been removed | 17:34 |
| ganso | gouthamr: wait | 17:34 |
| ganso | gouthamr: lol | 17:34 |
| gouthamr | ganso: yes, not for me. i'm still on the bugfix patch.. | 17:34 |
| gouthamr | :P | 17:34 |
| ganso | gouthamr: the task_state was already migration_success? | 17:34 |
| gouthamr | ganso: yeah.. the message is confusing.. even with notify gone, if someone uses that command after a migration has been completed, they will get that error message.. | 17:35 |
| ganso | gouthamr: I can add a different message for that task_state | 17:35 |
| ganso | gouthamr: and possibly for migration_error and None | 17:35 |
| gouthamr | ganso: yes.. icing on the cake.. | 17:35 |
| gouthamr | ganso: btw, my dictionary is formatted like this {'phase': 'replicating', 'state': 'healthy', 'estimated_completion_time': None, 'details': 'Volume move job monitoring transfer of data. ', 'total_progress': None} | 17:37 |
| gouthamr | ganso: sigh, see a bug. i should set total_progress to 0 if None. | 17:37 |
| ganso | gouthamr: exactly | 17:37 |
| *** kaisers_ has joined #openstack-manila | 17:37 | |
| ganso | gouthamr: I were going to hit the check I am adding right now | 17:37 |
| ganso | gouthamr: lol | 17:37 |
| gouthamr | ganso: still, don't expect 'current_file_progress'. :) | 17:38 |
| gouthamr | or path. | 17:38 |
| ganso | gouthamr: ya | 17:38 |
| *** eharney has joined #openstack-manila | 17:40 | |
| *** kaisers_ has quit IRC | 17:42 | |
| *** aovchinnikov has joined #openstack-manila | 17:48 | |
| *** zhugaoxiao has joined #openstack-manila | 17:53 | |
| *** Administrator__ has quit IRC | 17:55 | |
| *** eharney has quit IRC | 17:56 | |
| *** rraja has quit IRC | 18:02 | |
| *** eharney has joined #openstack-manila | 18:08 | |
| openstackgerrit | Alyson proposed openstack/manila: Fix HNAS Driver accepting invalid entries as EVS IP https://review.openstack.org/359222 | 18:08 |
| openstackgerrit | Rodrigo Barbieri proposed openstack/manila: Fix Share Migration improper behavior for drivers https://review.openstack.org/332267 | 18:22 |
| gouthamr | ganso: https://review.openstack.org/#/c/332267/57..58/manila/api/views/shares.py <-- those keys should always be returned | 18:28 |
| *** lpetrut has joined #openstack-manila | 18:33 | |
| ganso | gouthamr: why? | 18:40 |
| gouthamr | ganso: API sanity.. you don't want to selectively return different keys.. the response schema shouldn't change based off the situation | 18:41 |
| gouthamr | ganso: the client can figure out that the current_file_in_progress being None means that it can't have that information.. | 18:41 |
| ganso | gouthamr: ok, then I believe it is best to have only 'total_progress' | 18:42 |
| ganso | gouthamr: as we discussed the other day, we don't know what fields all drivers will be able to provide | 18:43 |
| *** timcl has quit IRC | 18:53 | |
| openstackgerrit | Rodrigo Barbieri proposed openstack/manila: Fix Share Migration improper behavior for drivers https://review.openstack.org/332267 | 18:54 |
| ganso | gouthamr: ^ | 18:55 |
| gouthamr | ganso: thanks | 18:55 |
| ganso | gouthamr: thank you for the feedback :) | 18:56 |
| *** AJaeger has joined #openstack-manila | 18:57 | |
| AJaeger | bswartz, could you review https://review.openstack.org/357924 ,please? That allows to move forward with api-ref work | 18:57 |
| *** dsariel has quit IRC | 19:02 | |
| AJaeger | thanks, bswartz ! | 19:06 |
| *** timcl has joined #openstack-manila | 19:11 | |
| *** AJaeger has left #openstack-manila | 19:15 | |
| *** pcaruana has quit IRC | 19:20 | |
| *** kaisers_ has joined #openstack-manila | 19:26 | |
| *** kaisers_ has quit IRC | 19:30 | |
| openstackgerrit | Merged openstack/manila: Get ready for os-api-ref sphinx theme change https://review.openstack.org/357924 | 19:33 |
| *** aovchinnikov has quit IRC | 19:35 | |
| *** lpetrut has quit IRC | 20:03 | |
| *** porrua has quit IRC | 20:04 | |
| *** akerr has quit IRC | 20:09 | |
| gouthamr | ganso: here? | 20:11 |
| ganso | gouthamr: yes | 20:11 |
| gouthamr | ganso: 2 quick questions for you | 20:12 |
| ganso | gouthamr: go ahead | 20:12 |
| gouthamr | ganso: so, if phase1 is done, why can't you use get-progress/ | 20:12 |
| ganso | gouthamr: because the jobs running in storage backends from where the progress is obtained is most probably finished | 20:13 |
| ganso | gouthamr: but I see the point you raise... that is should be total_progress = 100 in that case | 20:14 |
| *** alkhodos has quit IRC | 20:14 | |
| gouthamr | ganso: yes | 20:18 |
| gouthamr | ganso: maybe make the call and drivers should support it? | 20:19 |
| gouthamr | ganso: i don't see why they can't.. | 20:19 |
| ganso | gouthamr: yea, they may just return 'total_progress' = 100 themselves | 20:19 |
| gouthamr | ganso: +1 | 20:22 |
| gouthamr | ganso: another question.. is there a way to estimate the time remaining for data copy based off your current progress? | 20:23 |
| gouthamr | i was just wondering if it would be helpful to have two things in the API response.. to save administrators from becoming impatient (in bswartz's words) | 20:23 |
| ganso | gouthamr: HDS storage can, and with a reasonable amount of work we can support it in fallback too | 20:25 |
| gouthamr | ganso: i presume drivers can give you that, or return 'unknown' | 20:25 |
| gouthamr | ala windows software :P | 20:25 |
| ganso | gouthamr: windows software usually lie, it would be better if they returned 'unknown' lol | 20:26 |
| gouthamr | total progress: 0% estimated time remaining: unknown msg: go take a walk | 20:26 |
| ganso | gouthamr: I've seen: estimated time remaining: 2 hours. <5 minutes later>. Estimated time remaining: 10 minutes | 20:27 |
| gouthamr | ganso: yeah, you just got upgraded to first class. | 20:27 |
| gouthamr | :P | 20:28 |
| ganso | gouthamr: or worse, stays at 'estimated time remaining: 1 minute' to keep you close by while it takes half an hour | 20:28 |
| gouthamr | ganso: that'd be a pain.. but it's an 'estimate' and we have the license to be woefully off mark in cases | 20:29 |
| *** rhefner has quit IRC | 20:31 | |
| *** rhefner has joined #openstack-manila | 20:33 | |
| ganso | gouthamr: yes | 20:33 |
| ganso | gouthamr: still, I would rather return only 'total_progress' right now and work on that once we have more drivers on board | 20:34 |
| gouthamr | ganso: sure. makes sense | 20:34 |
| openstackgerrit | Alyson proposed openstack/manila: Fix HNAS Driver accepting invalid entries as EVS IP https://review.openstack.org/359222 | 20:41 |
| *** rhefner has quit IRC | 20:43 | |
| ganso | gouthamr: I'll update the patch later tonight | 20:45 |
| ganso | gouthamr: with the progress change for the phase1_done | 20:46 |
| *** rhefner has joined #openstack-manila | 20:48 | |
| *** cknight has quit IRC | 20:50 | |
| *** vbellur has quit IRC | 20:50 | |
| gouthamr | ganso: sure.. no rush. the latest patch LGTM | 20:51 |
| *** timcl has quit IRC | 20:56 | |
| *** gouthamr has quit IRC | 20:58 | |
| *** kaisers_ has joined #openstack-manila | 21:15 | |
| *** kaisers_ has quit IRC | 21:20 | |
| *** eharney has quit IRC | 21:27 | |
| *** gouthamr has joined #openstack-manila | 21:34 | |
| openstackgerrit | Merged openstack/manila-image-elements: Move other-requirements.txt to bindep.txt https://review.openstack.org/354924 | 21:42 |
| *** lpetrut has joined #openstack-manila | 22:24 | |
| *** lpetrut has quit IRC | 22:28 | |
| *** xyang1 has quit IRC | 22:36 | |
| *** vbellur has joined #openstack-manila | 22:41 | |
| *** chlong has quit IRC | 23:03 | |
| *** alyson_ has quit IRC | 23:03 | |
| openstackgerrit | Rodrigo Barbieri proposed openstack/manila: Share migration Newton improvements https://review.openstack.org/328431 | 23:49 |
| openstackgerrit | Rodrigo Barbieri proposed openstack/manila: Fix Share Migration improper behavior for drivers https://review.openstack.org/332267 | 23:49 |
| openstackgerrit | Rodrigo Barbieri proposed openstack/manila: Add share type change to Share Migration https://review.openstack.org/345142 | 23:49 |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!