Tuesday, 2017-09-05

*** wanghao has quit IRC00:54
*** wanghao has joined #openstack-mogan00:55
zhenguomorning mogan!01:05
zhenguoliusheng: can we format the server node with name and uuid as well?01:11
zhenguoliusheng: as we only support list node name01:11
zhenguoliusheng: admin have to go to ironic to see the uuid and name relationship01:12
liushengzhenguo: if so we need to call ironic01:12
zhenguoliusheng: can we?01:12
liushengzhenguo: and not sure if Ironic support duplicated name01:12
zhenguoliusheng: no01:12
zhenguoliusheng: we also use name in RP01:13
zhenguoliusheng: it can't be duplicated01:13
liushengzhenguo: maybe it is better to support that in server side ? wdyt ?01:13
zhenguoliusheng: you mean return the node name instead of uuid?01:13
liushengzhenguo: because usually, the API return is more exact and client side transform it to human readable01:14
zhenguoliusheng: yes01:15
zhenguoliusheng: we can change to use node name instead01:15
liushengzhenguo: use node name in where ?01:17
zhenguoliusheng: API01:17
liushengzhenguo: I mean maybe we can return node uuid in API, or both node name and node uuid01:17
zhenguoliusheng: which should be consistent with our node list output01:17
zhenguoliusheng: we just return the node uuid now01:18
zhenguoliusheng: on client side, seems you can't call ironic to get the name if you are not admin01:19
liushengzhenguo: I found we only return node name01:19
zhenguoliusheng: I mean server show01:19
liushengzhenguo: oh, sorry :(01:20
zhenguoliusheng: server show should be consistent with node list01:20
liushengzhenguo: hmmm, seems yes01:20
zhenguoliusheng: and aggregate node list01:20
zhenguoliusheng: I will make a change to replace it with name on API side01:21
zhenguoliusheng: seems need to change the scheduler return01:21
liushengzhenguo: I am afraid if it is a choice if we always use node name than node uuid, since may we will support other drivers, we need to ensure the node name is unique in other drivers01:22
zhenguoliusheng: that's a limitation of placement01:23
zhenguoliusheng: the rp name is unique01:23
liushengzhenguo: and need to sure it can support get a node with node name directly01:23
zhenguoliusheng: so if other driver's name is not unique they should make it unique01:23
liushengzhenguo: GET /xxx/node-name01:23
zhenguoliusheng: ironic support that01:23
liushengzhenguo: I mean we need ensure other drivers suppport that01:24
zhenguoliusheng: why01:24
liushengzhenguo: it is convenient to get a node info, if we always use node name than node uuid01:24
zhenguoliusheng: in fact, we don't use node name but the RP's name01:25
zhenguoliusheng: driver can decide uuid or name for the RP name01:25
liushengzhenguo: oh, yes01:25
zhenguoliusheng: RP name is always unique01:25
zhenguoliusheng: they can use uuid for Rp name01:26
liushengzhenguo: make sense01:28
openstackgerritMerged openstack/python-moganclient master: Format the output for server object  https://review.openstack.org/50051101:51
zhenguoliusheng: can you please help to review the up patches? we can't move along without your reviews :P03:13
liushengzhenguo: ok, reviewing03:15
openstackgerritZhenguo Niu proposed openstack/mogan master: Return node(name or uuid) with server instead of node_uuid  https://review.openstack.org/50066403:37
openstackgerritXinran WANG proposed openstack/mogan master: Use attach_interface during the server creation  https://review.openstack.org/50066804:00
Xinranzhenguo,  pls review this patch if you got time, sorry for the delay. I wasn't in beijing these days.  https://review.openstack.org/50066804:01
openstackgerritMerged openstack/mogan master: Rollback to the original status when server powering action failed  https://review.openstack.org/47686204:42
zhenguoXinran: thanks04:44
*** wanghao has quit IRC05:50
*** wanghao has joined #openstack-mogan05:51
*** wanghao_ has joined #openstack-mogan06:07
*** wanghao has quit IRC06:10
zhenguoliusheng: I just deleted you mogan project on RTD06:13
liushengzhenguo: ok, you need a new project ?06:14
zhenguoliusheng: I tried to use another name with my own repo, it built successfully06:14
liushengzhenguo: ok, feel free to test06:15
zhenguoliusheng: seems we don't have permissions to add webhook to openstack/mogan06:18
liushengzhenguo: seems yes, I have tried, but don't know how to add06:18
zhenguoliusheng: I got a warning that there's no permissions06:19
zhenguoliusheng: in fact, it will automatically added06:19
zhenguoliusheng: you can check you own repo's settings/webhook06:19
liushengzhenguo: I found that, but it cannot work06:20
zhenguoliusheng: http://mogan.readthedocs.io/en/latest/06:25
*** litao__ has joined #openstack-mogan06:26
liushengzhenguo: this looks good06:26
zhenguoliusheng: hah06:26
liushengzhenguo: what changes you made ?06:26
zhenguoliusheng: just changed to my repo06:26
liushengzhenguo: ... :D06:27
zhenguoliusheng: deleted yours and create a new one named mogan06:27
liushengzhenguo: strange06:27
*** wanghao has joined #openstack-mogan06:41
*** wanghao_ has quit IRC06:42
liushengzhenguo: don't need to save the partition info into db ?06:52
openstackgerritZhenguo Niu proposed openstack/mogan master: Return node(name or uuid) with server instead of node_uuid  https://review.openstack.org/50066406:54
zhenguoliusheng: not sure06:54
zhenguoliusheng: do you think we need to save that information06:54
liushengzhenguo: maybe user want to know, such as before rebuilding a server06:54
liushengzhenguo: also not sure, hah06:54
zhenguoliusheng: seems it's useful,06:55
zhenguoliusheng: which format do you like?06:55
liushengzhenguo: or maybe only visable by admins06:55
zhenguoliusheng: seems we should expose that to users, as it just set by them06:55
liushengzhenguo: seems the input format is good06:55
zhenguoliusheng: hah06:56
liushengzhenguo: yes06:56
zhenguoliusheng: we may support more advanced partitions when ironic side is done06:56
zhenguoliusheng: like LVM06:56
liushengzhenguo: ok, sounds good06:57
zhenguoliusheng: seems nova will same root_gb, ephemeral_gb, and swap_mb separatedly06:57
liushengzhenguo: same/save ?06:58
zhenguoliusheng: oh, save06:58
liushengzhenguo: hah, I think the benifit for separated saving is for the usage of retriving, like filter query07:00
liushengzhenguo: but seems don't need have this requirement07:00
zhenguoliusheng: yes07:00
zhenguoliusheng: seems we don't need that07:00
zhenguoliusheng: and also find ironic support to pass some root disk hints07:00
zhenguoliusheng: to choose which disk to be used as root07:00
liushengzhenguo: nova don't support that ?07:01
liushengzhenguo: I mean Nova + Ironic07:01
zhenguoliusheng: seems they can07:01
zhenguoliusheng: but actually, they don't know what disk they can choose, lol07:01
liushengzhenguo: oh, we can implement that later maybe in Q07:01
liushengzhenguo: lol07:01
zhenguoliusheng: I plan to pass something like (> root+ephemeral+swap)07:02
zhenguoliusheng: they can support such hints07:02
zhenguoliusheng: will add on driver side07:02
zhenguoliusheng: maybe a following up patch07:02
liushengzhenguo: ok, sounds like useful features07:03
liushengzhenguo: so you prefer to merge your partition patch before releasing and improve later, or add support to save the partition info ?07:05
zhenguoliusheng: before07:07
zhenguoliusheng: prior07:07
liushengzhenguo: ok, hah07:07
zhenguoliusheng: it's almost good, need to wait infra guys to land the project config parth07:08
liushengzhenguo: otherwise, it cannot pass gate ?07:08
zhenguoliusheng: yes07:08
zhenguoliusheng: as the default tempest partition image doesn't contain grub207:08
liushengzhenguo: oh, may need to ping some infra guys if this is urgent07:09
zhenguoliusheng: seems some one is coming07:10
liushengzhenguo: hah, cool07:10
zhenguoliusheng: you can also leave a +1 there07:11
liushengzhenguo: I have given07:11
zhenguoliusheng: oh, it seems lost after I updated the patch07:12
liushengzhenguo: oh, yes, given again07:13
zhenguoliusheng: http://mogan.readthedocs.io/projects/api-ref/en/latest/v1/index.html07:13
zhenguoliusheng: hah I managed to make it work07:13
liushengzhenguo: cooool!07:13
zhenguoliusheng: hah07:13
liushengzhenguo: how to make it work ?07:13
zhenguoliusheng: created a _static manually07:13
zhenguoliusheng: but the change is on my private repo07:14
liushengzhenguo: oh, maybe we can merge it now, once we are official project, we can get rid of that07:14
zhenguoliusheng: I can't make openstack/mogan work07:14
zhenguoliusheng: only build with my own repo successfuly for docs07:15
liushengzhenguo: it is strange, same code07:15
zhenguoliusheng: yes07:15
zhenguoliusheng: but it's ok, I can upadte if needed, as the webhook is also not work for openstack/mogan07:15
liushengzhenguo: yes, that is a way07:16
liushengzhenguo: how if we specified the partitions size as 0 ?07:17
zhenguoliusheng: I will check it with the image size or min_size first07:19
zhenguoliusheng: for partiton images07:19
liushengzhenguo: ok07:20
zhenguoliusheng: for whole disk image, we don't support to specify partitions07:20
zhenguoliusheng: ironic will check as well07:20
liushengzhenguo: ok07:21
zhenguoliusheng: please have a look at this https://review.openstack.org/#/c/500452/ as well07:29
liushengzhenguo: ok, sure07:29
*** wanghao has quit IRC07:57
*** wanghao has joined #openstack-mogan07:57
openstackgerritZhenguo Niu proposed openstack/mogan master: Add root disk partitions support  https://review.openstack.org/49903909:00
zhenguoliusheng: hi, do you like to keep the RTD theme for moganclient or chanage it to openstackdoctheme09:04
openstackgerritZhenguo Niu proposed openstack/mogan master: Add preserve_ephemeral support for rebuilding  https://review.openstack.org/50045209:08
liushengzhenguo: if it can work with openstackdoctheme in RTD, may 2ed choice is better09:24
liushengzhenguo: what do you think ?09:24
zhenguoliusheng: yes09:27
zhenguoliusheng: you can make the change09:28
zhenguoliusheng: and remember to add docs09:28
liushengzhenguo: let me try09:28
liushengzhenguo: yes, I am working on09:28
zhenguoliusheng: thanks09:28
liushengzhenguo: np :D09:28
*** wanghao has quit IRC09:29
*** wanghao has joined #openstack-mogan09:29
*** wanghao has quit IRC09:29
*** wanghao has joined #openstack-mogan09:30
*** wanghao has quit IRC09:30
*** wanghao has joined #openstack-mogan09:31
*** wanghao has quit IRC09:31
*** wanghao has joined #openstack-mogan09:31
*** wanghao has quit IRC09:32
*** wanghao has joined #openstack-mogan09:32
*** wanghao has quit IRC09:33
*** wanghao has joined #openstack-mogan09:33
*** wanghao has quit IRC09:33
*** wanghao has joined #openstack-mogan09:34
*** wanghao has quit IRC09:34
*** wanghao has joined #openstack-mogan09:35
openstackgerritXinran WANG proposed openstack/mogan master: Update port['device_id'] when creating server  https://review.openstack.org/50066810:02
zhenguoliusheng: and this https://review.openstack.org/#/c/498775/10:59
liushengzhenguo: ok11:01
*** litao__ has quit IRC12:13
*** shaohe_feng has quit IRC13:06
*** shaohe_feng has joined #openstack-mogan13:14
*** shaohe_feng has quit IRC13:33
*** shaohe_feng has joined #openstack-mogan13:33
openstackgerritMerged openstack/mogan master: [DOC] Add states and transitons diagram  https://review.openstack.org/47129314:22
openstackgerritMerged openstack/mogan master: Add root disk partitions docs  https://review.openstack.org/50031614:23
openstackgerritMerged openstack/mogan master: Return node(name or uuid) with server instead of node_uuid  https://review.openstack.org/50066415:01
openstackgerritYaguo Zhou proposed openstack/mogan master: Replace DbMigrationError with DBMigrationError  https://review.openstack.org/50087015:04
*** lin_yang has joined #openstack-mogan15:18
*** harlowja has joined #openstack-mogan17:08
*** lin_yang has quit IRC19:01
*** harlowja has quit IRC21:13

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