Tuesday, 2016-08-30

« Monday, 2016-08-29 Index Wednesday, 2016-08-31 »
*** gouthamr_ has joined #openstack-manila00:04
*** gouthamr has quit IRC00:08
*** mtanino has quit IRC00:16
*** Suyi_ has quit IRC00:16
*** akerr has joined #openstack-manila00:19
openstackgerritRodrigo Barbieri proposed openstack/manila: Share migration Newton improvements  https://review.openstack.org/32843100:52
*** akerr has quit IRC00:55
*** kaisers_ has quit IRC01:14
*** yangyapeng has joined #openstack-manila01:18
openstackgerritRodrigo Barbieri proposed openstack/manila: Share migration Newton improvements  https://review.openstack.org/32843101:24
*** wangqun has joined #openstack-manila01:32
*** ociuhandu has joined #openstack-manila01:32
*** ociuhandu has quit IRC01:34
*** akerr has joined #openstack-manila01:59
*** akerr is now known as akerr_away02:19
*** akerr_away is now known as akerr02:19
*** akerr is now known as akerr_away02:20
*** akerr_away is now known as akerr02:20
*** akerr is now known as akerr_away02:20
*** akerr_away is now known as akerr02:20
*** akerr is now known as akerr_away02:21
*** akerr_away is now known as akerr02:21
*** akerr is now known as akerr_away02:22
openstackgerritRodrigo Barbieri proposed openstack/manila: [ZFSonLinux] Add share migration support  https://review.openstack.org/35341702:22
*** akerr_away is now known as akerr02:22
*** akerr is now known as akerr_away02:23
*** akerr_away is now known as akerr02:23
*** gouthamr_ has quit IRC02:30
*** magic has joined #openstack-manila02:32
*** magic is now known as Guest3343102:32
*** xiaoyang has quit IRC02:35
*** draynium has quit IRC02:45
*** draynium has joined #openstack-manila02:48
*** ociuhandu has joined #openstack-manila02:53
*** akerr_ has joined #openstack-manila03:02
*** kaisers_ has joined #openstack-manila03:02
openstackgerritPeter Wang proposed openstack/manila: VNX: Use job for NFS share creation  https://review.openstack.org/35956703:05
*** akerr has quit IRC03:06
*** ociuhandu has quit IRC03:07
*** kaisers_ has quit IRC03:08
*** akerr_ has quit IRC03:23
openstackgerritNam Nguyen Hoai proposed openstack/manila: TrivialFix: Remove cfg import unused  https://review.openstack.org/36096003:32
openstackgerritMerged openstack/manila: Implement replication support in huawei driver  https://review.openstack.org/33158604:05
*** ganso has quit IRC04:25
*** senk has joined #openstack-manila04:34
*** kaisers_ has joined #openstack-manila04:52
*** kaisers_ has quit IRC04:57
*** sandanar has joined #openstack-manila05:47
*** nkrinner_afk is now known as nkrinner05:48
*** kaisers_ has joined #openstack-manila05:56
*** rhefner has quit IRC06:20
*** pcaruana has joined #openstack-manila06:26
*** lpetrut has joined #openstack-manila06:36
*** ubijtsa has joined #openstack-manila06:39
*** ubijtsa is now known as Guest5456806:39
*** Guest54568 has quit IRC06:42
*** Guest54568 has joined #openstack-manila06:42
*** Guest54568 is now known as assassin06:43
*** rraja has joined #openstack-manila06:58
*** dsariel has joined #openstack-manila07:09
*** lpetrut has quit IRC07:21
*** akapil has joined #openstack-manila07:32
openstackgerritgecong proposed openstack/manila: Change assertTrue(isinstance()) by optimal assert  https://review.openstack.org/36267107:39
*** kaisers1 has quit IRC07:50
*** kaisers has quit IRC07:52
*** senk has quit IRC07:52
*** kaisers has joined #openstack-manila07:53
*** kaisers1 has joined #openstack-manila07:53
*** vponomaryov has joined #openstack-manila08:07
*** kaisers_ has quit IRC08:07
*** kaisers_ has joined #openstack-manila08:08
openstackgerritgecong proposed openstack/manila: Change assertTrue(isinstance()) by optimal assert  https://review.openstack.org/36267108:11
*** magic has joined #openstack-manila08:18
*** magic is now known as Guest9997308:19
*** Guest33431 has quit IRC08:22
*** aovchinnikov has joined #openstack-manila08:48
*** akapil has quit IRC08:58
*** akapil has joined #openstack-manila08:59
*** lpetrut has joined #openstack-manila09:06
openstackgerritAlexey Ovchinnikov proposed openstack/manila: [DNM] Tests for scenario tests  https://review.openstack.org/34023609:26
*** nkrinner has quit IRC09:29
*** lpetrut1 has joined #openstack-manila09:34
*** lpetrut has quit IRC09:34
*** lpetrut1 is now known as lpetrut09:34
*** nkrinner has joined #openstack-manila09:35
*** kaisers_ has quit IRC09:48
*** wangqun has quit IRC10:08
*** yangyapeng has quit IRC10:21
openstackgerritAlexey Ovchinnikov proposed openstack/manila: [DNM] Tests for scenario tests  https://review.openstack.org/34023610:34
openstackgerritBhagyashri Shewale proposed openstack/python-manilaclient: Replace functions 'Dict.get' and 'del' with 'Dict.pop'  https://review.openstack.org/36185810:37
*** akapil has quit IRC10:44
*** timcl has joined #openstack-manila10:54
*** akapil has joined #openstack-manila10:58
*** akapil_ has joined #openstack-manila10:59
*** akapil has quit IRC10:59
*** ganso has joined #openstack-manila11:18
openstackgerritzhongjun proposed openstack/manila-specs: Add spec for enable IPv6 in manila  https://review.openstack.org/36278611:27
openstackgerritValeriy Ponomaryov proposed openstack/manila: [Tempest] Run tempest with the only plugin - manila's  https://review.openstack.org/33837111:33
gansovponomaryov: welcome back11:33
vponomaryovganso: hey, Hello Rodrigo11:34
gansovponomaryov: Hello Valeriy. How was your vacation?11:34
vponomaryovganso: great )11:34
gansovponomaryov: =D11:34
vponomaryovganso: Now I have loong list of things to look at )11:35
gansovponomaryov: yes ^_^11:36
*** kaisers_ has joined #openstack-manila11:37
*** timcl has quit IRC11:39
*** kaisers_ has quit IRC11:41
*** tpsilva has joined #openstack-manila11:48
*** yangyapeng has joined #openstack-manila11:53
*** sandanar has quit IRC12:04
*** timcl has joined #openstack-manila12:05
tbarronnow the manila channel looks right again12:08
tbarronvponomaryov: welcome back!12:08
vponomaryovtbarron: hello Tom ))12:08
openstackgerritRodrigo Barbieri proposed openstack/manila: Share migration Newton improvements  https://review.openstack.org/32843112:11
openstackgerritRodrigo Barbieri proposed openstack/manila: Fix Share Migration improper behavior for drivers  https://review.openstack.org/33226712:11
openstackgerritAlexey Ovchinnikov proposed openstack/manila: [DNM] Tests for scenario tests  https://review.openstack.org/34023612:15
*** gouthamr has joined #openstack-manila12:17
openstackgerritRodrigo Barbieri proposed openstack/manila: Share migration Newton improvements  https://review.openstack.org/32843112:38
openstackgerritRodrigo Barbieri proposed openstack/manila: Add share type change to Share Migration  https://review.openstack.org/34514212:38
*** akapil_ has quit IRC12:41
*** rooneym has joined #openstack-manila12:41
*** gouthamr_ has joined #openstack-manila12:49
*** gouthamr has quit IRC12:51
*** akerr has joined #openstack-manila12:53
openstackgerritMerged openstack/manila-ui: Fix tables layout  https://review.openstack.org/36058812:55
*** yangyapeng has quit IRC12:57
*** rraja has quit IRC13:00
*** xyang1 has joined #openstack-manila13:02
*** david-lyle has quit IRC13:10
openstackgerritAlexey Ovchinnikov proposed openstack/manila: [DNM] Tests for scenario tests  https://review.openstack.org/34023613:16
*** david-lyle has joined #openstack-manila13:18
openstackgerritValeriy Ponomaryov proposed openstack/manila-ui: Add share replication support  https://review.openstack.org/22638913:21
*** dustins has joined #openstack-manila13:26
*** rraja has joined #openstack-manila13:43
openstackgerritAlexey Khodos proposed openstack/manila: Nexenta: adding share drivers for NexentaStor  https://review.openstack.org/30928613:43
openstackgerritAlexey Khodos proposed openstack/manila: Nexenta: adding share drivers for NexentaStor  https://review.openstack.org/30928613:45
*** baojg has joined #openstack-manila13:46
gansobswartz, xyang1, markstur, vponomaryov, tbarron, gouthamr_ : Hello, https://review.openstack.org/#/c/328431/ is waiting for review, passed all jenkins jobs. Please take a look when you have some time13:51
gouthamr_ganso: sure.. would like to see the ZFSOnLinux migration patch pass on that.. it currently needs some more work..13:52
*** gouthamr_ is now known as gouthamr13:53
gouthamrganso: also, were you going to add 'task_state' to get_progress?13:54
gansogouthamr: nope. I preferred to stay with the error message which informs the user of what the share migration state is13:55
bswartzganso: isn't this more important? https://review.openstack.org/#/c/345125/13:57
gansobswartz: that one requires https://review.openstack.org/#/c/328437/ and https://review.openstack.org/#/c/345142/ to merge first, which require the one I linked you13:58
bswartzI see13:58
*** faiz89 has joined #openstack-manila14:01
*** eharney has joined #openstack-manila14:02
openstackgerritMerged openstack/manila: Fix Share Migration improper behavior for drivers  https://review.openstack.org/33226714:03
bswartzganso how can I sqaush 328431 together with 353417?14:03
bswartzI'm trying to test all this stuff together14:04
*** cknight has joined #openstack-manila14:04
gansobswartz: cherry pick 353417 on top of 328431 and update migration_check_compatiblity accordingly14:05
openstackgerritValeriy Ponomaryov proposed openstack/manila-ui: Add share replication support  https://review.openstack.org/22638914:05
gansobswartz: and if you are running tempest, add run_optimized_migration_tests = True to your tempest.conf14:06
bswartzganso: okay I'll try that14:06
bswartzgah, still merge conflict14:08
bswartzvponomaryov: I want to rebase 353417 on 32843114:08
bswartzvponomaryov: any issues with that?14:09
openstackgerritAlexey Khodos proposed openstack/manila: Nexenta: adding share drivers for NexentaStor  https://review.openstack.org/30928614:09
vponomaryovI applied it on top of current master with "driver fixes" commit that has just been merged14:09
vponomaryovI should do it on top of that other commit?14:09
*** rraja has quit IRC14:09
vponomaryovbswartz/ ganso?14:10
gansoI'll try that rebase here, just a sec14:10
bswartzvponomaryov: I think it needs rebased on top of "Share migration Newton improvements"14:10
vponomaryovganso: can you answer question?14:10
gansovponomaryov: ideally, it is best, because you will make your ZFS driver implementation compatible with preserve-metadata and non-disruptive parameters14:11
gansovponomaryov: else it will just assume your driver cannot do any of those14:11
gansobswartz: conflict is in post_test_hook14:12
gansobswartz: a single line14:12
openstackgerritAlexey Ovchinnikov proposed openstack/manila: [DNM] Tests for scenario tests  https://review.openstack.org/34023614:12
bswartzganso: I know, but who is going to push the rebase? I'm offering to do it but I don't want to clobber anybody14:13
vponomaryovI am going14:13
gansobswartz: I'm on it14:13
bswartzganso: ty14:13
openstackgerritRodrigo Barbieri proposed openstack/manila: [ZFSonLinux] Add share migration support  https://review.openstack.org/35341714:13
vponomaryovganso; just apply rebase on top of master for 43114:14
vponomaryovbest coop ever ))14:14
gansovponomaryov: I am confused, didn't bswartz just say he wanted to rebase on top of 417?14:14
bswartzganso: you rebased on 34514214:16
bswartzwas that correct?14:16
*** baojg has quit IRC14:16
gansobswartz: it is similarto 41714:16
gansobswartz: does not affect any driver interface14:16
bswartzokay14:16
* bswartz head spins with too many changes14:17
bswartzI think this will work I'm running tests now14:17
*** baojg has joined #openstack-manila14:18
*** rraja has joined #openstack-manila14:22
*** Yogi1 has joined #openstack-manila14:22
*** yangyapeng has joined #openstack-manila14:25
openstackgerritRodrigo Barbieri proposed openstack/manila: [Dummy driver] Add share migration support  https://review.openstack.org/34785314:26
*** baojg has quit IRC14:28
openstackgerritValeriy Ponomaryov proposed openstack/manila: [Dummy driver] Add share migration support  https://review.openstack.org/34785314:33
*** ociuhandu has joined #openstack-manila14:33
bswartzganso: ERROR: The resource could not be found.14:33
bswartz:-(14:34
openstackgerritValeriy Ponomaryov proposed openstack/manila: [Dummy driver] Add share migration support  https://review.openstack.org/34785314:34
*** nkrinner has quit IRC14:36
*** kaisers_ has joined #openstack-manila14:37
bswartzganso: I'm getting a HTTP 404 from migration-start14:37
*** nkrinner has joined #openstack-manila14:39
*** aovchinnikov has quit IRC14:40
*** david-lyle has quit IRC14:43
*** nkrinner has quit IRC14:46
*** david-lyle has joined #openstack-manila14:48
openstackgerritValeriy Ponomaryov proposed openstack/manila: [ZFSonLinux] Add share migration support  https://review.openstack.org/35341714:58
bswartzvponomaryov: have you had any luck testing latest migration patch w/ ZFS?14:59
bswartzI think I must be doing something wrong14:59
*** nkrinner has joined #openstack-manila14:59
vponomaryovbswartz: it hangs in "migrating" state for me14:59
bswartzwhy would the migration-start API give me a 404?15:00
vponomaryovbswartz: latest PS has just fixes for rebase ganso made15:00
vponomaryovbswartz: when server does not support such API15:00
vponomaryovbswartz: have you restarted manila-api?15:00
vponomaryovbswartz: have you set up proper microversion?15:01
bswartzmy workflow is to download the patch before running stack.sh15:01
bswartzso no restart should be required15:01
*** david-lyle has quit IRC15:02
*** yangyapeng has quit IRC15:04
vponomaryovbswartz: manilaclient is updated as well?15:05
vponomaryovbswartz: so, just tested, it worked15:06
bswartzwith your latest change?15:08
bswartzthat's great15:08
bswartzI will test shortly15:08
vponomaryovyes15:08
bswartzgo ahead and put your reviews on ganso's patches15:08
vponomaryov1,5k changes ))15:09
vponomaryov+ 500 second15:09
vponomaryovwill not be fast15:09
*** akerr has quit IRC15:10
*** akerr has joined #openstack-manila15:11
*** nkrinner is now known as nkrinner_afk15:11
bswartzwell if you see any problems point them out as soon as you see them15:11
gansobswartz: that was probably your manilaclient15:11
ganso(I just got back)15:12
bswartzganso: what client should I use?15:12
vponomaryovhttps://review.openstack.org/#/c/328437/15:13
gansobswartz: https://review.openstack.org/34512515:13
vponomaryovany ^15:13
bswartzganso: what's what I have15:15
bswartzI wonder if it's not getting used for some reason15:15
*** kaisers_ has quit IRC15:19
*** david-lyle has joined #openstack-manila15:22
bswartzokay my patch download is what was failing15:22
bswartzI have no idea why...15:23
*** sandanar has joined #openstack-manila15:23
bswartzganso: I started the migration but when I get progress is say no migration in progress15:25
gansobswartz: what is the share's status and task_state?15:27
bswartzstatus available / task_state migration_error15:28
gansobswartz: it needs to be in migrating status and migration_driver_in_progress task_state to be able to obtain progress15:28
bswartzk this might be a subpar user experience15:29
bswartzif I start a migration and that succeeds, I'm going to follow up with a get progresse15:29
gansobswartz: the error message is correct, it seems to me you are not doing any migration to obtain progress15:29
*** kaisers_ has joined #openstack-manila15:29
bswartzeven if an error occurs, it makes sense for the get progress to tell me that15:29
bswartzotherwise it's unclear what API to call to query the progress15:30
*** senk has joined #openstack-manila15:30
gansobswartz: if an error occurs, the error message will be in the logs and task_state will be migration_error, there is no place to store that error message right now. Maybe when we have the jobs table we can store it there and include it in migration_get_progress15:31
gansobswartz: shouldn't this put on hold until Ocata? https://review.openstack.org/#/c/226389/15:40
*** david-lyle has quit IRC15:40
vponomaryovganso: why it should be on hold?15:42
gansovponomaryov: manila-ui FF was last week15:42
*** faiz89 has quit IRC15:42
gansovponomaryov: if it merges now, it will be included for RC1 version of manila-ui15:42
vponomaryovganso: it should have been merged long time ago15:42
gansovponomaryov: indeed, it should have15:43
openstackgerritValeriy Ponomaryov proposed openstack/manila: [Dummy driver] Add share migration support  https://review.openstack.org/34785315:45
*** lpetrut has quit IRC15:48
*** david-lyle has joined #openstack-manila15:50
*** timcl1 has joined #openstack-manila15:51
*** timcl has quit IRC15:54
vponomaryovbswartz, ganso: zfsonlinux passed migration tests -> http://logs.openstack.org/17/353417/13/check/gate-manila-tempest-dsvm-postgres-zfsonlinux/8177733/console.html#_2016-08-30_15_51_15_00871016:15
*** ociuhandu has quit IRC16:15
*** lpetrut has joined #openstack-manila16:20
*** david-lyle has quit IRC16:21
*** david-lyle has joined #openstack-manila16:23
*** kaisers_ has quit IRC16:28
gansovponomaryov:  \o/16:28
gansogouthamr: ping16:31
*** david-lyle_ has joined #openstack-manila16:36
*** timcl1 has quit IRC16:39
*** david-lyle has quit IRC16:39
bswartzaha16:45
bswartzI think I found out why my test doesn't work >_<16:45
bswartzganso, vponomaryov: this is what I'm currently getting http://paste.openstack.org/show/564860/16:48
vponomaryovbswartz: use --writable False16:49
bswartzdoh!16:49
vponomaryovoptional argument for command with migration16:49
vponomaryovit is True by default16:49
bswartzgood catch vponomaryov16:49
vponomaryovalso, you may want to disable fallback migration16:49
vponomaryovwith --skip_optimized_migration True16:50
bswartzvponomaryov: why didn't you support writable migration with your original implementation?16:50
vponomaryovbswartz: "writable" means we can write to share and be sure all changes are applied16:50
vponomaryovbswartz: it is false for case of zfsonlinux driver16:50
vponomaryovbswartz; write while migrating16:51
bswartzokay16:51
bswartzbut we *could* support it16:51
vponomaryovbswartz: no16:51
bswartzjust need to take an extra snapshot and transfer it during the cutover16:51
gansovponomaryov: you cannot disable fallback migration. Skip_optimized_migration = True will skip your driver's implementation and run fallback16:51
vponomaryovbswartz: it will be endless loop16:51
vponomaryovganso: right, with False16:52
bswartzvponomaryov: no, the design of the feature is that when the migration is completed, it's disruptive16:52
vponomaryovganso: but as you can see, it is confusing right now ))16:52
vponomaryovfallback, optimized...16:52
bswartzso after the disruption you can send one last snapshot and be sure you got all the data16:52
vponomaryovneed to use driver-based and generic migrations16:52
gansovponomaryov: you can use a parameter that prevents fallback from running, such as preserve-metadata or nondisruptable16:52
bswartzactually I agree with half of that -- I don't like "optimized"16:52
bswartzI would prefer "driver assisted" and "fallback" migration types16:53
vponomaryovbswartz: I do not see real value in making such feature, we either make sure that "everything" was ported or do not support it16:53
vponomaryovin the former case it is endless loop16:53
bswartz"optimized" implies it's somehow better than the fallback migration16:53
vponomaryovin case we write constantly16:53
gansobswartz: gouthamr doesn't like fallback... I kinda like it but I understand his point... need something better else I think it is fine to stick with it16:53
gansobswartz: it usually is better16:54
bswartzvponomaryov: I can be done -- but I agree it's additional code we don't need to write today16:54
gansobswartz: at least, it is supposed to be16:54
vponomaryovbswartz: even if we had time16:54
vponomaryovbswartz: we should assume constant writing16:54
bswartzganso: it's different -- usually better, but not required -- therefore "optimized" is misleading16:54
vponomaryovbswartz: it is endless, cannot support16:54
bswartzvponomaryov: what you're forgetting is that the share is unexported when you call migration complete, so you can't continue writing after that16:55
bswartzin the time between the migration-complete API call and the migration actually finishing, there is no access to the share16:56
bswartzwe can use that window of disruption to ensure no data loss16:56
*** erlon has quit IRC16:57
gouthamr"fallback" sounds like dev language to me..17:02
bswartzhttp://paste.openstack.org/show/564864/17:03
gouthamrlike imho, we should be saying 'driver-assisted' vs 'host-assisted' data copy :P17:03
openstackgerritLucian Petrut proposed openstack/manila: Windows SMB: implement 'update_access' method  https://review.openstack.org/35485517:03
*** faiz89 has joined #openstack-manila17:05
openstackgerritAlexey Ovchinnikov proposed openstack/manila: [DNM] Tests for scenario tests  https://review.openstack.org/34023617:08
*** Suyi_ has joined #openstack-manila17:14
gansogouthamr: not bad17:17
gansogouthamr: can we talk about migration_get_progress change?17:18
*** sandanar_ has joined #openstack-manila17:21
gouthamrganso: yep...17:21
gansogouthamr: ok so, I got your point that the current approach is not good for scripts that keep checking progress17:22
gansogouthamr: Also, could you please clarify your comment in https://review.openstack.org/#/c/328431/50/manila_tempest_tests/tests/api/admin/test_migration_negative.py@174 ?17:23
*** sandanar has quit IRC17:24
*** pcaruana has quit IRC17:25
gouthamrganso: yes.. we have negative tests where requesting an API with a version less than the version in which it was changed is validated..17:25
gansogouthamr: but if we are dropping support, why do we need to have a test for that?17:26
gouthamrganso: dropping support to?17:26
gansogouthamr: everything migration-related17:26
gansogouthamr: only 2.22 and above will work17:26
gouthamrganso: i'm talking about preserve_metadata, writable, nondisruptive17:26
gouthamrganso: those are available only 2.22 and beyond?17:26
gansogouthamr: yes17:27
gouthamrganso: also, allowing 'unsetting' task state should only be allowed 2.22 and beyond?17:27
* bswartz fails again17:27
bswartzMANILA_ZFSONLINUX_USE_SSH=True17:27
gansogouthamr: yes, I am considering that as well17:27
gouthamrganso: so those can be added as negative tests..17:27
gansogouthamr: the API will just reject17:28
gansogouthamr: because it is below the minimum supported17:28
gouthamrganso: would be a simple ddt test, and you expect HTTP40017:28
gouthamrganso: no, you're microversioning the body params.. and not the API..17:28
gansogouthamr: but I am changing the API as well to accept only 2.22 and beyond17:28
gouthamrganso: oh. wait, migration won't work at all under 2.22?17:29
gansogouthamr: no, the API will reject17:29
gansogouthamr: say the action is not found17:29
gouthamrganso: o.O i'm stupid if i overlooked that in my review and tests..17:29
gansogouthamr: I thought you suggested that when you said to remove all backwards compatibility17:30
gouthamrganso: that's a huge change..17:30
gansogouthamr: so it is as if migration never existed before17:30
gouthamrganso: i was suggesting you break microversion support entirely because we seem to have a loose definition of that for experimental APIs anyway.. i meant, if a user uses 2.15, give him all the features of 2.22.. because 2.15 wouldn't know that you dropped support for notify..17:31
gansogouthamr: I was doing exactly that before17:32
gansogouthamr: main problem was testing manilaclient17:32
gansogouthamr: I don't remember what it was exactly, but only the latest could be used17:33
gansogouthamr: shell could not be microversioned with different parameters17:34
gouthamrganso: yes, i guess we only test latest in the client..17:34
gansogouthamr: so, the user had to have a real older version of python-manilaclient17:34
gouthamrganso: we've done that before.. os-some_action to some_action17:34
gansogouthamr: and I had already accomplished that, just removing the commit17:34
gansogouthamr: yes, but the CLI command must not be the same17:35
gansogouthamr: or in fact, must not have different parameters17:35
gouthamrganso: oh.. yes.. these changes can only be done inside the client17:35
gouthamrganso: not shell..17:35
gouthamrganso: so the command is the same, but since everything goes through the client, you can fence off invalid req parameters there..?17:36
gansogouthamr: if the user has the latest client, it cannot invoke a previous microversion method with an older-named parameter17:37
*** rraja has quit IRC17:37
gansogouthamr: the client would have to know that, like for example, skip_optimized_migration is force_host_copy in previous versions17:37
gansogouthamr: so he would use skip_optimized_migration in shell, and then the client would send force_host_copy17:38
gouthamrganso: oh.. so that's why this API only makes sense in 2.22 and beyond?17:40
gansogouthamr: right now, yes17:41
bswartzganso: new error "ERROR: Invalid share: Migration progress of share d7bded87-3d00-4bec-aaa3-52dffca6797c cannot be obtained at this moment."17:42
gansobswartz: what is your status and task_state at the moment?17:43
bswartzstatusmigrating,    task_statemigration_driver_phase1_done17:43
gouthamrganso: that's what i mean ^17:43
bswartzwtf spacing17:43
bswartz status migrating, task_state migration_driver_phase1_done17:44
bswartzganso it was a transient thing17:44
gouthamrganso: i am okay with 2.22 being the minimum microversion... can we call that out in the commit message and the reno?17:44
bswartzeventually it returned "ERROR: Invalid share: Migration of share d7bded87-3d00-4bec-aaa3-52dffca6797c has already completed first phase."17:44
gansobswartz: you should have received ""Migration of share %s has already completed first phase." message instead of that one17:44
bswartzso it's weird that there's a race condition where you can get the first error17:45
gansobswartz: what I am doing right now is that I am changing all those "error" messages to a standard dictionary response that includes the task_state17:45
bswartzthe share should instantly go into a state where progress is reported as zero when the start completes17:45
bswartzand when it's done it should just say done instead of "error"17:46
*** lpetrut has quit IRC17:46
gansobswartz: so instead of the message you got, you will get {'total_progress': 100, 'task_state': 'migration_driver_phase1_done'}17:46
gansobswartz: as zero?17:46
gansobswartz: I am hardcoding it as 100 right now17:46
gansobswartz: oh, when migration-start completes... sorry I misread 'when migration completes'17:47
gouthamrvponomaryov: still around by any chance?17:48
bswartzganso: yeah there seems to be a race after you start the migration when you get progress17:49
bswartzif you get progress too fast you get the error I got17:49
gansobswartz: it is not really an error, that's why I am changing that now17:49
gansobswartz: it is just informative17:49
*** gouthamr has quit IRC17:54
*** gouthamr has joined #openstack-manila17:54
*** senk has quit IRC18:02
*** sandanar_ has quit IRC18:03
*** akerr is now known as akerr_away18:04
*** akerr_away is now known as akerr18:04
gansogouthamr: ping18:20
gouthamrganso: pong18:20
gansobswartz, gouthamr: are you ok with changing "skip_optimized_migration" to "skip_driver_assisted_migration" ?18:20
gansobswartz, gouthamr: or "force_host_assisted_migration"18:21
bswartzganso: yes18:23
gansobswartz: which one?18:23
gouthamrganso: i like "force_host_assisted_migration"18:23
bswartzganso: however I was wondering if we could convert the CLL flags that are --xxx True/False to jut --xxx with no boolean, so the presence of the flag determines the behavior18:25
gansobswartz: depends on the default value18:25
gansobswartz: writable is true by default18:26
gansobswartz: so it needs --writable False18:26
gansobswartz: the only ones that can do that is --nondisruptive and --skip_optimized_migration18:26
openstackgerritJay Mehta proposed openstack/manila: HPE 3PAR driver pool support  https://review.openstack.org/32955218:26
gansobswartz: but I don't like the idea of half parameters behaving one way and the other half behaving differently18:26
bswartzganso: I'm in favor of changing them all18:31
bswartzI agree consistency is important18:31
bswartzdo you have examples of ones we can't change because they've non-experimental?18:31
gansobswartz: yes but what I said is that some cannot be just --writable, such as 'writable'18:32
gansobswartz: it is not related to being experimental18:32
gansobswartz: it is because of their default value18:32
gansobswartz: --writable defaults to True18:32
gansobswartz: so you need the parameter to disable it18:32
gansobswartz: or, change it to --nonwritable18:32
bswartzganso: the flag would be --readonly, and it would default to falkse18:33
bswartzfalse18:33
gansobswartz: that too18:33
gansobswartz: how about preserve-metadata ?18:33
gouthamr(--preserve-metadata (true) / --preserve-metadata=True) vs (--preserve-metadata=False)18:35
gouthamr(--writable / --writable=True vs --writable=False)18:35
gansogouthamr: ?18:35
gouthamr(--skip-optimized-migration / --skip-optimized-migration=True) vs (--skip-optimized-migration=False)18:36
gouthamrsorry, was just enumerating the options..18:36
gouthamri didn't understand the confusion18:36
gansooh18:37
gansoso, we need to invert preserve-metadata18:37
ganso--writable would change to --readonly as bswartz suggested18:37
*** faiz89 has quit IRC18:38
bswartzganso: --no-preserve-metadata? --degrade-metadata? --modify-metadata?18:40
gansobswartz: those look terrible18:40
gansobswartz: lol18:40
gansobswartz: we've had --lossy discussed in the midcycle18:40
gansobswartz: but --lossy is bad18:40
bswartzbetter than --corrupt-metadata or --destroy-metadata18:40
gouthamr--readonly? It's strange that the adminstrator will tell you that he/she wants the share to be readonly..18:40
gouthamrlike that's saying i want this migration only if you can do it by removing all current write access on the share18:41
*** mtanino has joined #openstack-manila18:41
ganso--disregard-metadata ?18:41
gansogouthamr: yes it is a bit weird18:41
gansogouthamr: writable is True by default and admin says "ok if you cannot do it writable, then do it non-writable"18:42
bswartzganso: https://review.openstack.org/#/c/22638918:42
bswartzganso: well what we want to communicate is that the metadata will be modified in the same way that tar/untar would modify it18:42
bswartzso it's not really "lost" or "disregarded" it's just degraded to minimum posix owners/acls18:43
*** lpetrut has joined #openstack-manila18:45
bswartzganso: --posix-metadata-only ?18:46
gansobswartz: :\18:47
bswartzI don't see how --no-preserve-metadata is any different than --preserve-metadata False18:47
*** dsariel has quit IRC18:47
gansobswartz: then we can stick with '--preserve-metadata False' :P18:49
bswartzyeah but --xxx True/False is just bad UI design18:49
bswartzI can think of almost no software that does CLI flags that wya18:49
bswartzway18:49
*** Suyi_1 has joined #openstack-manila18:50
gansoOpenStack has some CLIs that are like that18:50
bswartzboolean flags are always simply flags with no following parameter18:50
*** Suyi_ has quit IRC18:52
bswartzganso: it's not a huge deal for me18:52
bswartzagain it's something we could change later if needed18:52
bswartzjust wanted to express my opinion18:53
bswartzganso: btw ZFS migrations are working reliably for me now that I fixed my env18:53
*** Yogi1 has quit IRC18:56
*** erlon has joined #openstack-manila18:59
*** david-lyle_ is now known as david-lyle19:05
*** lpetrut has quit IRC19:10
*** faiz89 has joined #openstack-manila19:21
*** akshai has joined #openstack-manila19:53
*** harlowja has quit IRC19:58
*** akerr has quit IRC20:16
*** dsariel has joined #openstack-manila20:18
*** FL1SK has quit IRC20:30
tbarronganso: bswartz Downstream I need to backport https://review.openstack.org/#/c/363118/ - would be great to get it into stable/mitaka upstream if it's acceptable.20:38
*** faiz89 has quit IRC20:52
*** cknight has quit IRC20:53
*** lpetrut has joined #openstack-manila20:54
*** dustins has quit IRC21:04
*** faiz89 has joined #openstack-manila21:10
*** faiz89 is now known as faiz89_away21:10
*** faiz89_away is now known as faiz8921:10
*** harlowja has joined #openstack-manila21:17
*** gouthamr has quit IRC21:19
*** faiz89 has quit IRC21:20
*** rhefner has joined #openstack-manila21:23
*** dustins has joined #openstack-manila21:26
*** ganso has quit IRC21:35
*** dsariel has quit IRC21:36
*** ganso has joined #openstack-manila21:39
*** eharney has quit IRC21:48
*** dustins has quit IRC21:50
*** cknight has joined #openstack-manila21:58
*** cknight1 has joined #openstack-manila22:00
*** cknight has quit IRC22:02
*** alyson_ has quit IRC22:03
*** lpetrut has quit IRC22:12
*** cknight1 has quit IRC22:24
*** FL1SK has joined #openstack-manila22:39
openstackgerritJay Mehta proposed openstack/manila: HPE 3PAR driver pool support  https://review.openstack.org/32955223:37
*** tpsilva has quit IRC23:44
*** rooneym has quit IRC23:49
openstackgerritRodrigo Barbieri proposed openstack/manila: Share migration Newton improvements  https://review.openstack.org/32843123:56
*** akerr has joined #openstack-manila23:58
« Monday, 2016-08-29 Index Wednesday, 2016-08-31 »

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