Wednesday, 2016-11-30

openstackgerritMerged openstack/manila-ui: Adds Manila-UI Devstack plugin  https://review.openstack.org/38885900:31
*** nherciu has quit IRC00:44
*** tommylikehu_ has joined #openstack-manila00:48
*** zhonghua_ has joined #openstack-manila00:52
*** mtanino has quit IRC00:59
*** zhonghua_ has quit IRC01:13
panatltbarron: i think i am facing this issue .. https://review.openstack.org/#/c/403651/01:19
tbarronpanatl: I won't be online much tonight, but is the context here still with the very simple two-line devstack deployment?01:23
panatlyes01:23
tbarronpanatl: so maybe apply that patch and restart the manila services?01:24
panatlunless there is some good official devstack doc .. :)01:24
tbarronpanatl: :)01:24
panatlyeah i will try that patch ...thanks!01:24
tbarronpanatl: I'll try it out myself, but I got diverted by family stuff tonight.  I'll be online tomorrow though if you want to sync up again.01:25
panatlyep, let's sync-up tomorrow01:25
panatlit is late for me too ... will start ./stack.sh . and grab a bottle .. good night and thanks!01:26
*** gcb has joined #openstack-manila01:50
*** zhonghua has quit IRC01:51
*** zhonghua has joined #openstack-manila01:51
openstackgerritCedric Zhuang proposed openstack/manila: Add "update_access" interface support for VNX.  https://review.openstack.org/39540401:57
*** kaisers_ has joined #openstack-manila02:07
*** kaisers has quit IRC02:10
*** zengyingzhe__ has joined #openstack-manila02:29
*** tuanluong has joined #openstack-manila02:31
*** zengyingzhe_ has quit IRC02:31
openstackgerritHa Van Tu proposed openstack/manila: [hacking] Ensure not to use LOG.warn  https://review.openstack.org/37715503:15
*** zhonghua has quit IRC03:46
*** zhonghua has joined #openstack-manila03:46
*** tommylikehu_ has quit IRC03:51
*** kaisers_ has quit IRC04:53
*** cknight has joined #openstack-manila05:00
*** tommylikehu_ has joined #openstack-manila05:07
*** catinthe_ has quit IRC05:19
*** catintheroof has joined #openstack-manila05:21
*** senk has joined #openstack-manila05:24
*** shausy has joined #openstack-manila05:26
*** catintheroof has quit IRC05:26
*** kaisers has joined #openstack-manila05:53
*** gcb has quit IRC05:54
*** kaisers has quit IRC06:02
*** gouthamr has joined #openstack-manila06:06
*** gcb has joined #openstack-manila06:07
*** cknight has quit IRC06:13
openstackgerritzhangyanxian proposed openstack/manila: Fix usage of LOG mismatch in the code  https://review.openstack.org/40346506:14
openstackgerritzhangyanxian proposed openstack/manila: Fix usage of LOG mismatch in the code  https://review.openstack.org/40346506:32
*** igajsin has quit IRC06:35
openstackgerritHa Van Tu proposed openstack/manila: [hacking] Ensure not to use LOG.warn  https://review.openstack.org/37715506:36
*** igajsin has joined #openstack-manila06:36
openstackgerritzhangyanxian proposed openstack/manila: LOG marker mismatch in the code  https://review.openstack.org/40457006:36
openstackgerritHa Van Tu proposed openstack/manila: [hacking] Ensure not to use LOG.warn  https://review.openstack.org/37715506:37
openstackgerritzhangyanxian proposed openstack/manila: LOG marker mismatch in the code  https://review.openstack.org/40457006:38
openstackgerritPony Chou proposed openstack/manila: Add QNAP Manila Driver  https://review.openstack.org/39470306:41
openstackgerritzhangyanxian proposed openstack/manila: LOG marker mismatch in the code  https://review.openstack.org/40457006:41
*** sandanar has joined #openstack-manila06:42
*** tuanluong has quit IRC06:42
*** senk has quit IRC06:50
openstackgerritzhangyanxian proposed openstack/manila: LOG marker mismatch in the code  https://review.openstack.org/40457006:58
*** kaisers has joined #openstack-manila06:58
*** nkrinner_afk is now known as nkrinner07:04
*** kaisers has quit IRC07:06
*** hoonetorg has quit IRC07:21
*** hoonetorg has joined #openstack-manila07:21
*** hoonetorg has quit IRC07:24
*** hoonetorg has joined #openstack-manila07:24
tommylikehu_40458807:25
*** mkoderer has joined #openstack-manila07:30
*** jprovazn has joined #openstack-manila07:31
*** rraja has joined #openstack-manila07:47
*** dsariel has joined #openstack-manila07:51
*** gcb has quit IRC07:51
*** gcb has joined #openstack-manila07:53
*** xinyanzhang has joined #openstack-manila07:54
*** senk has joined #openstack-manila07:57
*** kaisers has joined #openstack-manila08:03
*** kaisers has quit IRC08:13
openstackgerritPony Chou proposed openstack/manila: Add QNAP Manila Driver  https://review.openstack.org/39470308:16
*** makowals_ has joined #openstack-manila08:30
*** makowals_ has quit IRC08:31
*** makowals has quit IRC08:33
*** makowals has joined #openstack-manila08:44
*** senk has quit IRC09:03
*** akapil has joined #openstack-manila09:05
*** zhonghua_ has joined #openstack-manila09:09
*** kaisers has joined #openstack-manila09:10
*** lpetrut has joined #openstack-manila09:10
*** a-pugachev has joined #openstack-manila09:11
*** kaisers has quit IRC09:17
tommylikehu_tbarron, vponomaryov, bswartz, markstur, could we talk about the detail on IPv6 spec?09:39
tommylikehu_obtain ip version by parsing export location's ip address VS  write this in  export location metadata (perhaps)09:39
tommylikehu_I am not sure which one should be addressed in next patch09:41
*** ganso has joined #openstack-manila09:49
*** zhonghua_ has quit IRC09:51
*** tommylikehu_ has quit IRC09:51
*** tommylikehu_ has joined #openstack-manila09:51
*** tommylikehu_ has quit IRC09:51
*** tommylikehu_ has joined #openstack-manila09:52
*** lpetrut has quit IRC09:55
*** tommylikehu_ has quit IRC09:56
*** lpetrut has joined #openstack-manila10:03
*** kaisers has joined #openstack-manila10:14
*** hoonetorg has quit IRC10:15
*** andreaf has quit IRC10:17
*** sapcc-bot has quit IRC10:17
*** carthaca_1 has quit IRC10:17
*** sapcc-bot has joined #openstack-manila10:17
*** david_1 has joined #openstack-manila10:17
*** tpatzig_ has joined #openstack-manila10:17
*** databus23_ has joined #openstack-manila10:17
*** carthaca_ has joined #openstack-manila10:17
*** mkoderer_ has joined #openstack-manila10:17
*** dgonzalez_ has joined #openstack-manila10:17
*** tommy_ has joined #openstack-manila10:17
*** tommy_ is now known as Guest2455610:18
*** andreaf has joined #openstack-manila10:18
*** mkoderer___ has joined #openstack-manila10:18
*** mkoderer__ has quit IRC10:19
*** logan- has quit IRC10:19
*** darrenc has quit IRC10:19
*** carthaca_ has quit IRC10:19
*** tpatzig_ has quit IRC10:19
*** Guest24556 has quit IRC10:19
*** dgonzalez_ has quit IRC10:19
*** mkoderer_ has quit IRC10:19
*** databus23_ has quit IRC10:19
*** david_1 has quit IRC10:19
*** sticker has quit IRC10:19
*** bswartz has quit IRC10:19
*** lpetrut has quit IRC10:20
*** darrenc has joined #openstack-manila10:20
*** logan- has joined #openstack-manila10:20
*** carthaca_ has joined #openstack-manila10:20
*** kaisers has quit IRC10:21
*** hoonetorg has joined #openstack-manila10:24
*** vkmc has quit IRC10:25
*** makowals has quit IRC10:25
*** makowals has joined #openstack-manila10:26
*** sticker has joined #openstack-manila10:27
*** vkmc has joined #openstack-manila10:27
*** pcaruana has joined #openstack-manila10:35
openstackgerritPony Chou proposed openstack/manila: Add QNAP Manila Driver  https://review.openstack.org/39470310:38
openstackgerritMerged openstack/manila: Fix devstack smb configuration outside ubuntu  https://review.openstack.org/40443310:41
openstackgerritMerged openstack/manila: [hacking] Ensure not to use LOG.warn  https://review.openstack.org/37715510:48
*** kaisers has joined #openstack-manila10:51
*** tpsilva has joined #openstack-manila10:51
*** lpetrut has joined #openstack-manila10:55
*** kaisers has quit IRC10:59
*** lpetrut has quit IRC11:01
*** akapil has quit IRC11:01
gouthamrganso: ping11:06
gansogouthamr: pong11:07
gouthamrganso: regarding the comments on https://review.openstack.org/#/c/399049/8/specs/ocata/fix-and-improve-access-rules.rst ..11:07
gouthamrganso: i wasn't proposing the cast_to_readonly parameter to be added. does it appear like i am?11:08
gouthamrganso: i'm confused how to reword that..11:08
gansogouthamr: I think it would be perfect if you just state that the column will be added and what it will be used for, and that there will be a condition in update_access to accomplish that. Is there any reason the share_manager needs to be involved in this?11:09
gouthamrganso: hmmm... you mean strip out the discussion regarding the alternative11:10
*** kaisers1 has quit IRC11:10
gansogouthamr: what discussion regarding the alternative? I may have missed that11:10
gansogouthamr: I didn't find anything related to it in the alternatives section11:11
gouthamrganso: alright, i'll move it to alternatives where it probably makes more sense.. i get the confusion now11:13
*** digvijay2016 has joined #openstack-manila11:21
*** kaisers has joined #openstack-manila11:26
*** akapil has joined #openstack-manila11:27
*** nherciu has joined #openstack-manila11:30
openstackgerritMerged openstack/manila: [Dummy driver] Add possibility to set delays for driver methods  https://review.openstack.org/40033111:37
*** jprovazn has quit IRC11:40
openstackgerritAlyson proposed openstack/manila: Add support for manage/unmanage snapshots in HNAS driver  https://review.openstack.org/40125311:44
tbarronmkoderer: ping - when you have a chance I have some questions about external storage backends and DHSS=True where I understand you have real-world experience11:48
tbarronpanatl: you must be hitting a different issue than https://review.openstack.org/#/c/403651/ since that local.conf deploys generic driver with DHSS=True11:52
tbarronpanatl: I spun up a devstack with that local.conf and can try to repro your issue if you have steps - maybe put them in a pastebin11:53
*** kaisers_ has joined #openstack-manila11:56
*** porrua has joined #openstack-manila12:03
*** tinwood has quit IRC12:04
*** kaisers_ has quit IRC12:04
*** tinwood has joined #openstack-manila12:05
mkoderertbarron: pong12:11
tbarronmkoderer: are you deploying hardware storage backends with DHSS=True drivers?12:12
mkoderertbarron: yes we do12:12
mkoderertbarron: we have it live in production12:13
tbarronmkoderer: have you ever figured how to do this when a data interface on the backend only has an access-connection (no vlan/vxlan) to the switch connecting it into the OpenStack cloud?12:13
mkoderertbarron: means you have VLAN or VXLAN in Neutron but flat networking in the storage network?12:14
tbarronmkoderer: yes, there is a lab switch with its own vlan dedicated to "the storage network".12:15
mkoderertbarron: so we don't have such a setup but it should work12:15
mkodereryou need to use a different network driver for such a setup12:16
mkodererbut I never tested that12:16
tbarronmkoderer: but the h/w backend just has an access connection to the switch, so all traffic is in the lab storage vlan12:16
tbarronmkoderer: guess I'd need to use the singlenetwork network backend rather than neutron backend ...12:17
mkoderertbarron: yeah basically you don't need to have any VLAN id passed to the storage backend12:18
tbarronmkoderer: do you know what network one would use for the share network in that case?12:18
vponomaryovtbarron: it depends on network plugin you use12:19
vponomaryovtbarron: if you use standalone one, then you configure net details via manial config explicitly12:20
vponomaryovtbarron: without any neutron-specific IDs12:20
tbarronmkoderer: in the neutron plugin case I'd use the private net for compute instance, have a floating ip, etc.12:20
tbarronvponomaryov: yeah, I think I get that, is the share-net just ignored when I supply it then?12:20
mkoderertbarron: as vponomaryov wrote - so if you use NeutronSingleNetworkPlugin there must be on neutron net representing your network12:21
vponomaryovtbarron: supply what and where?12:21
tbarronvponomaryov: i'm still required to supply a shared net in the create-share command, right?12:21
tbarrona share-network12:21
* tbarron gets easily confused sometimes12:22
vponomaryovtbarron: why "shareD" net?12:22
vponomaryovtbarron: it is not "shared"12:22
vponomaryovtbarron: it belongs to share12:22
vponomaryovtbarron: or serves it12:22
tbarronvponomaryov: I misspoke.  s/shared/share/12:22
vponomaryovtbarron: and yes, share-network is required always in case of DHSS=True mode12:23
mkoderertbarron: for DHSS=true you need to specify a share net12:23
vponomaryovtbarron: but network info can be taken in 2 ways12:23
vponomaryovtbarron: the way of getting network info depends on network plugin you use12:24
tbarronvponomaryov: I'm confused about the share-network that one supplies when creating a share with NeutronSingleNetworkPlugiin.12:24
vponomaryovtbarron: share-network is used for more than one purpose12:24
vponomaryovtbarron: not only providing INFO12:25
vponomaryovtbarron: also for tracking share-servers12:25
vponomaryovtbarron: all shares of one share-network are located on one share server12:25
vponomaryovtbarron: and vice versa12:25
tbarronvponomaryov: ack12:25
vponomaryovtbarron: then what is confusing?12:25
mkoderertbarron: are you confused because  of the need to create a "empty" share-network without any ids?12:26
tbarronmkoderer: yes.  didn't really know I can do that.  I thought I have to refer to network-id and subnet-id from neutron.12:28
tbarronvponomaryov: ^^^12:28
tbarronmkoderer: vponomaryov my h/w backend is using a network and IP addresses outside neutron's management, reachable via default route out the external network.12:29
*** catintheroof has joined #openstack-manila12:29
tbarronmkoderer: vponomaryov works fine with DHSS=False12:30
vponomaryovtbarron: then use StandAlone network plugin12:30
vponomaryovtbarron: for DHSS=True mode12:30
vponomaryovtbarron: it is exactly the use case of it12:30
tbarronvponomaryov: and create a share-network that I can refer to, right?12:30
vponomaryovtbarron: yes12:30
tbarronwhat do I use for network ID and subnetwork ID when creating the share network?12:31
*** makowals_ has joined #openstack-manila12:31
vponomaryovtbarron: empty share networks for each set of combined shares on single share server12:31
vponomaryovtbarron: nothing12:31
vponomaryovtbarron: you just create empty share-network12:31
vponomaryovtbarron: a name it in some fancy way )12:32
vponomaryovs/a/and/12:32
*** catinthe_ has joined #openstack-manila12:32
tbarronvponomaryov: ok, that was what was confusiing me, thanks!12:32
tbarronmkoderer: thanks to you too12:32
vponomaryovtbarron: has confusion been annihilated completely? or left something else?12:32
tbarronvponomaryov: nature abhors a vacuum.  I'll get some more confusion soon :)12:33
*** catintheroof has quit IRC12:34
*** makowals has quit IRC12:34
* tbarron hadn't noticed that --neutron-net-id and --neutron-subnet-id are **optional** arguments to 'manila share-network-create'. D'oh ...12:34
vponomaryovtbarron: then fill yourself with some Zen ))12:35
tbarronvponomaryov: there's nothing there whenever I sit still and look12:35
tbarronvponomaryov: well, 'I' don't "sit still and look"12:36
vponomaryovtbarron: there should be something to tease ))12:36
vponomaryovganso: JFYI: https://bugs.launchpad.net/manila/+bug/164609712:37
openstackLaunchpad bug 1646097 in Manila "Generic driver does not have admin export location with DHSS=False mode" [High,New] - Assigned to Valeriy Ponomaryov (vponomaryov)12:37
gansovponomaryov: so all you gotta do is create a new key with "admin_ip"?12:39
vponomaryovganso: it is one of the ways to fix bug, yes12:40
gansovponomaryov: cool12:40
vponomaryovganso: but it is not enough to fix general connection problem )12:40
gansovponomaryov: :\12:40
mkoderervponomaryov: ganso: can we merge https://review.openstack.org/#/c/404118/12:40
mkoderertrivial one12:40
gansomkoderer: sure12:41
vponomaryovmkoderer: maybe there is something else that is unused?12:41
vponomaryovmkoderer: you sure it is the only orphaned DB method?12:41
mkoderervponomaryov: I checked for volume_ and didn't find anything12:42
vponomaryovmkoderer: it is not the only criterion12:42
vponomaryovmkoderer: =)12:42
mkoderervponomaryov: I know :P this one I found during a review ;)12:43
mkoderervponomaryov: if you know a fancy regex I am open for it :P12:43
*** jprovazn has joined #openstack-manila12:43
vponomaryovmkoderer: its ok, lots of such small steps will make us reach the ideal12:43
*** makowals has joined #openstack-manila12:44
vponomaryov^_^12:44
*** dsariel has quit IRC12:44
a-pugachevHi, can someone take a look at https://review.openstack.org/#/c/404167/ ?12:45
*** makowals_ has quit IRC12:45
a-pugachevit’s simple one, to tidy up coverage report12:45
openstackgerritValeriy Ponomaryov proposed openstack/manila: [DNM] test 8  https://review.openstack.org/40474912:48
*** digvijay2016 has quit IRC12:50
tbarrona-pugachev: i'll try running coverage with your patch in a bit12:54
*** ociuhandu has quit IRC12:55
*** gcb has quit IRC12:58
openstackgerritValeriy Ponomaryov proposed openstack/manila: [DNM] test 9  https://review.openstack.org/40475712:59
*** jcsp has quit IRC13:00
*** ociuhandu has joined #openstack-manila13:00
*** kaisers_ has joined #openstack-manila13:01
*** ociuhandu has quit IRC13:07
*** kaisers_ has quit IRC13:09
*** catintheroof has joined #openstack-manila13:12
*** dsariel has joined #openstack-manila13:16
*** catinthe_ has quit IRC13:16
*** gouthamr has quit IRC13:17
*** shausy has quit IRC13:24
*** arnewiebalck has joined #openstack-manila13:27
*** xyang1 has joined #openstack-manila13:30
*** kaisers_ has joined #openstack-manila13:33
*** kaisers_ has quit IRC13:39
*** porrua has quit IRC13:53
*** StraubTW has joined #openstack-manila14:04
*** kaisers_ has joined #openstack-manila14:05
openstackgerritArne Wiebalck proposed openstack/manila-ui: Add access key column to rules table  https://review.openstack.org/39371514:07
panatltbaaron: Steps are ..14:07
*** tongli has joined #openstack-manila14:07
*** porrua has joined #openstack-manila14:08
panatltbaaron: neutron net-list14:08
panatltbaaron: neutron subnet-list14:09
panatltbaaron: manila share-network-create --neutron-net-id <NET_ID> --neutron-subnet-id <SUBNET_ID> --name manila_share14:09
*** kaisers_ has quit IRC14:09
panatltbaaron: manila share-network-list14:09
panatltbaaron: tbaaron: manila create --name devstack_share --share-network <SHARE_NET_ID> NFS 114:09
panatltbaaron:  now after this ... it hangs since it try to ssh to that NOVA VM ... but ssh login is brokern ...14:09
rrajapanatl: tbaaron or tbarron?14:10
tbarronrraja: :)14:10
panatltbaaron: I am able to create share .. after editing /etc/manila/manila.conf .. adding service_instance_password = manila14:10
*** nherciu has quit IRC14:10
panatltbaaron: and comnmenting private .. public key lines14:11
*** nherciu has joined #openstack-manila14:11
tbarronpanatl: so you are really saying that we have a doc bug?  that the two-line local.conf yields the wrong /etc/manila/manila.conf?  or perhaps that we should change the plugin to make the twoliine config work?14:12
*** senk has joined #openstack-manila14:12
tbarronpanatl: and rraja is right, for my irc to ping me you have to use 'tbarron'14:13
*** bswartz has joined #openstack-manila14:13
panatltbaaron: not sure ... depends .. i think we should let this work in ssh key mode as default ... instead of listing password in conf file14:13
panatlhaa sorry .. tbarron14:14
*** ianychoi has quit IRC14:14
tbarronpanatl: no problem :)14:14
panatli will leave it to core ... to decide  .. do we need this by default ork with SSH or change the prop .. and add that manila / manila .. user/pass14:15
panatli will certanly go for SSH to work by default .. that is better security14:16
*** eharney has joined #openstack-manila14:17
bswartzanyone seen tommylikehu this morning?14:18
tbarronpanatl: if you file a launchpad bug we'll remember to fix the issue one way or another14:19
panatltbarron: yep .. will do .. will also update doc page with the notes ... for any newbie .. like me :)14:20
*** akapil has quit IRC14:25
*** akapil has joined #openstack-manila14:26
xyang1ganso: ping14:27
*** gcb has joined #openstack-manila14:33
*** cknight has joined #openstack-manila14:33
*** sage__ has quit IRC14:35
*** kaisers_ has joined #openstack-manila14:38
tbarronpanatl: excellent, if you'd post the LP bug here that would be great14:38
*** mtanino has joined #openstack-manila14:43
panatlhttps://bugs.launchpad.net/manila/+bug/164613514:44
openstackLaunchpad bug 1646135 in Manila "ssh key login broken to share VM" [Undecided,New]14:44
*** kaisers_ has quit IRC14:44
tbarronpanatl: ty14:45
gansoxyang1: pong14:46
xyang1ganso: hi, all CI's have passed here: https://review.openstack.org/#/c/403444/  can you please take a look?14:46
gansoxyang1: sure14:47
xyang1ganso: thanks!14:47
vponomaryovpanatl: added answer to your bugreport14:52
vponomaryovpanatl: it works in Ci14:52
vponomaryovpanatl: so, my guess, there is something wrogn with setup of your SSH keys14:52
vponomaryovpanatl: either different user is used for keys or problem with key bits, something like this14:53
*** dustins has joined #openstack-manila14:53
vponomaryovpanatl: also, such bug, where you say that SSH is broken using key SHOULD include complete info about keys you use14:54
vponomaryovpanatl: service logs you get, etc14:54
*** chlong has joined #openstack-manila14:55
vponomaryovtbarron: ^14:55
tbarronvponomaryov: CI doesn't use that simple two line local.conf file https://github.com/openstack/manila/blob/master/devstack/README.rst14:55
tbarronvponomaryov: ^^ the README makes it look like that's all one has to do14:56
vponomaryovtbarron: have you tried two-liner yourself?14:56
tbarronvponomaryov: was going to do that in a few minutes14:56
*** alyson_ has joined #openstack-manila14:56
vponomaryovtbarron: two-liner means keys were not predefined14:57
tbarronvponomaryov: right14:57
tbarronvponomaryov: devstack will prompt you for passwords with the two liner, but probably not for manila user's password and won't setup keys, right?14:58
panatlvponomaryov: keys are created by stack process ... completly new empty 14 and 16 ubuntu14:58
panatleven the stack user is created by devstack/toosl/create-stacku**.sh script14:58
*** sage__ has joined #openstack-manila14:59
*** dustins has quit IRC15:01
openstackgerritMerged openstack/manila: Remove unused function in db api  https://review.openstack.org/40411815:03
*** eharney has quit IRC15:04
*** adrianofr_ has joined #openstack-manila15:05
*** nherciu has quit IRC15:05
*** nherciu has joined #openstack-manila15:06
*** Guest69699 has quit IRC15:06
*** Guest69699 has joined #openstack-manila15:06
*** dustins has joined #openstack-manila15:15
*** sandanar has quit IRC15:18
gansotbarron, bswartz, vponomaryov, cknight, toabctl, xyang1: Could you please take a look at this clean backport patch when you have some time? https://review.openstack.org/404700/ Thanks in advance15:21
*** senk has quit IRC15:27
*** eharney has joined #openstack-manila15:28
gansovponomaryov: I can't find anything related to migration in the logs of your latest test patch?15:28
gansovponomaryov: s/?/.15:28
*** Guest69699 is now known as assassin15:29
*** gouthamr has joined #openstack-manila15:35
*** gcb has quit IRC15:40
*** kaisers_ has joined #openstack-manila15:41
*** darrenc has quit IRC15:41
*** darrenc has joined #openstack-manila15:42
*** gouthamr has quit IRC16:02
*** chlong has quit IRC16:03
vponomaryovganso: which one is latest for you?16:05
vponomaryovganso: the most valueable commit with problem is here -> https://review.openstack.org/#/c/404757/16:05
vponomaryovganso: where, as I can see, mount operation hangs if we create some other share concurrently16:06
vponomaryovganso: and share service stops picking up messages from message bus16:06
*** rraja has quit IRC16:07
gansovponomaryov: the highest numbered? xD16:07
vponomaryovand fix works if we run it in single thread -> https://review.openstack.org/#/c/404394/16:07
gansovponomaryov: so this is a concurrency problem now... this is weird16:07
vponomaryovganso: migration works if we revert devstack's change16:07
vponomaryovganso: so, it is somehow related to the work of floating IPs16:08
*** cknight has quit IRC16:08
vponomaryovganso: can you say from top of your mind -> should we change access rules configuration when we change config of service VM?16:09
gansovponomaryov: please be more specific, what access rules what are you talking about, what for, and what config change of service VM?16:09
gansovponomaryov: I am lacking some more context16:10
vponomaryovganso: for data service16:10
*** Yogi1 has joined #openstack-manila16:10
vponomaryovganso: that mounts shares16:10
vponomaryovganso: and requires access rules16:10
*** ianychoi has joined #openstack-manila16:10
vponomaryovganso: service VM that is nova VM and configured to be used by generic driver in DHSS=False mode16:10
*** kaisers_ has quit IRC16:11
*** miriam_ has joined #openstack-manila16:16
gansovponomaryov: data service will allow access to the ip necessary to mount the share, which we assume there is stable connectivity. What config change could happen?16:16
vponomaryovganso; hey, it is my question ))16:18
*** chlong has joined #openstack-manila16:18
openstackgerritMerged openstack/manila: Move EMC drivers to dell_emc folder  https://review.openstack.org/40344416:18
gansovponomaryov: lol I don't see what the problem is16:18
gansovponomaryov: you said "when we change config of service VM"16:19
gansovponomaryov: so I am unaware of config changes you are talking about16:19
vponomaryovganso: yes, we changed network of service port16:19
gansovponomaryov: during migration?16:21
*** miriam_ has quit IRC16:21
gansovponomaryov: source and destination shares are being provided by same share server?16:21
*** cknight has joined #openstack-manila16:21
vponomaryovganso: Rodrigo )) configuration is changed compared to old runs16:22
vponomaryovganso: when feature was designed16:22
vponomaryovganso: not during migration ))16:22
gansovponomaryov: then there should be any problem, access rule should be the same always... what you could is that you could use the floating ip network like 172.24.4.0/2416:23
gansovponomaryov: should *not be any problem16:23
vponomaryovganso: no, usage of external network does not trigger NATing behavior16:23
vponomaryovganso: only direct usageof "floating IP" feature16:24
gansovponomaryov: so you cannot allow_access to 172.24.4.0/24 ?16:24
openstackgerritHelen Walsh proposed openstack/manila: VMAX manila plugin - Support for VMAX in Manila  https://review.openstack.org/40485916:25
vponomaryovganso: your statement is ambiguous16:25
vponomaryovganso: you mean use in access rule?16:25
gansovponomaryov:16:26
gansovponomaryov: yes16:26
gansovponomaryov: I see "mount -vt nfs 172.24.5.42"16:26
gansovponomaryov: so what if you use data_node_service_ip = 172.24.5.0/2416:26
gansovponomaryov: we are sure that in all runs the share will be accessible from a floating ip, we don't need to know exactly which one, we know they change16:27
vponomaryovganso: then, I would, say, name of option is incorrect16:27
vponomaryovganso: name assumes single IP16:27
vponomaryovganso: not range16:28
vponomaryovganso: only source doe says it can be range16:28
vponomaryov/doe/code/16:28
gansovponomaryov: same for "manila access-allow ip 172.24.5.0/24"16:28
vponomaryov^_^16:28
vponomaryovgreat defence!16:28
tbarron0.0.0.0/016:28
vponomaryov=)16:28
gansovponomaryov, tbarron: Tom would rather allow share access to the whole universe16:29
vponomaryovtbarron: honest people have nothing to hide? ))16:30
tbarron:)16:31
* tbarron hidess16:31
vponomaryovganso: found one more evidence16:32
vponomaryovganso: lock "generic_driver_mounts_FOO" is not released16:32
vponomaryovganso: never16:32
gansovponomaryov: :O16:33
gansovponomaryov: deadlock!16:33
vponomaryovyes )16:33
*** ianychoi has quit IRC16:33
vponomaryovwait, no16:33
vponomaryovnot dead loc16:33
vponomaryovendless lock16:33
vponomaryovthere is no second lock )16:33
gansovponomaryov: then what is the damage of that lock not being released?16:34
vponomaryovganso: it uses "server ID"16:34
vponomaryovganso: now guess how many server ID we have in DHSS=False mode )16:35
vponomaryovganso: hint, it is less than 2 and bigger than 0 ))16:35
gansovponomaryov: just 116:35
gansovponomaryov: obviously16:35
vponomaryovganso: yes! ))16:35
gansovponomaryov: sohuld be a million dollar question16:35
vponomaryovganso: so, left to get know why somethings hangs there16:36
vponomaryovit is bad that we do not have logs of all shell commands16:36
gansovponomaryov: root-wrap usually prints what it is invoking16:37
gansovponomaryov: but not the output16:37
vponomaryovganso: it is the most useful thing, the output16:38
vponomaryovganso: commands we know from source code )16:38
gansovponomaryov: yes lol16:38
*** kaisers_ has joined #openstack-manila17:08
*** dsariel has quit IRC17:09
*** catintheroof has quit IRC17:11
*** catintheroof has joined #openstack-manila17:12
*** nkrinner is now known as nkrinner_afk17:13
*** kaisers_ has quit IRC17:15
*** chlong has quit IRC17:21
*** pcaruana has quit IRC17:23
*** senk has joined #openstack-manila17:24
*** tongli has quit IRC17:30
*** chlong has joined #openstack-manila17:34
*** nherciu has quit IRC17:48
*** akapil_ has joined #openstack-manila17:58
*** akapil has quit IRC18:01
*** akapil_ has quit IRC18:02
*** a-pugachev has quit IRC18:06
*** kaisers_ has joined #openstack-manila18:12
openstackgerritVictoria Martinez de la Cruz proposed openstack/manila: Decouple Manila UI from Manila Devstack plugin  https://review.openstack.org/38885518:12
*** senk has quit IRC18:17
*** kaisers_ has quit IRC18:19
openstackgerritValeriy Ponomaryov proposed openstack/manila: [DNM] test 10  https://review.openstack.org/40491318:26
openstackgerritVictoria Martinez de la Cruz proposed openstack/manila-ui: Moves OPENSTACK_MANILA_SETTINGS TO local_settings.d/  https://review.openstack.org/39792618:27
*** ianychoi has joined #openstack-manila18:29
bswartztbarron: I responded to your PMs from yesterday18:31
tbarronbswartz: k, thanks, just read & that makes sense18:34
tbarronbswartz: i talked to vponomaryov and mkoderer about by situation this morning and probably i need to try to use the StandaloneNetworkPlugin.18:48
tbarronmkoderer hasn't exercised it in his setting but in principle it should work.18:49
bswartztbarron: what are you expecting to achieve with this configuration?18:55
tbarroni want to be able to have the backend spin up share servers so that I can create shares with DHSS=True18:56
bswartztbarron: you want to configure netapp cdot with share servers and connect from a client on the other side of a router?18:56
bswartztbarron: right but what then?18:56
bswartzmanual testing? tempest tests? scenario tests?18:57
tbarronbswartz: all of the above, but starting with #118:57
tbarron#1 testing scenarios of vms on private nets with floating ips being able to mount18:57
bswartzusing the standalone network plugin will require that you manage the pool of IPs somehow18:58
bswartzfor example mulitple test runs in parallel will cause issues unless they use non-overlapping pools of IPs18:58
tbarronI'll specify a pool from the addrs that have been reserved for the netapp to use18:58
bswartzeven if you serialize test runs, you need to keep the pool of IPs clear and probably include some custom logic to cleanup after failed test runs18:59
tbarronit/manila should in theory be able to handle concurrent requests from tenants on difft private nets18:59
bswartztbarron: yes but I'm worried about the possibility of more than one manila managing the same box (which would be common in an automated tempest test environment)18:59
tbarronbswartz: ack19:00
tbarronbswartz: i'm not really aiming for a CI system, just for being able to validate basics and19:00
bswartzk19:01
tbarronbswartz: then later being able to repro customer issues (well, approximate)19:01
tbarronapproximate b/c I would expect most customers would dedicate a switch and a netapp to their openstack deployment, whereas in this lab19:01
tbarronwe share the netapp among multiple openstacks and use a switch that is for the lab, not for a single openstack/neutron19:02
bswartzright19:03
* tbarron actually expects to find customers who are using their one netapp (or whatever) for vmware as well as for openstack.19:03
* tbarron but he doesn't want to think about that kind of thing ...19:04
bswartztbarron: use cases like that are why are added the singlesvm mode19:10
bswartzhonestly if you use the multisvm mode and share the cluster with something like vmware I'd expect bad things to happen19:11
tbarronbswartz: +119:11
*** kaisers_ has joined #openstack-manila19:16
*** kaisers_ has quit IRC19:25
vponomaryovganso, bswartz: revert-back of devstack's change that broke DHSS=False mode of Generic driver is not working anymore19:31
vponomaryovganso, bswartz: proof: http://logs.openstack.org/70/402970/1/check/gate-manila-tempest-dsvm-generic-no-share-servers-ubuntu-xenial-nv/235eb34/console.html#_2016-11-30_18_57_21_86878919:31
vponomaryovganso, bswartz: it means, existing fix is ok compared to old approach WE used. It is something else that was merged to either nova, neutron or cinder19:32
bswartzvponomaryov: that's not surprising at all -- I was more surprised when you said the revert fixed it19:33
vponomaryovbswartz: it did19:33
bswartzthis means we're really dealing with 2 separate bugs19:33
vponomaryovseveral days ago19:33
*** mkoderer has quit IRC19:33
vponomaryovbswartz: another proof, I cannot reproduce bug with projects pulled up on 24th of November19:34
vponomaryovhaving latest manila + fix19:35
bswartzso what does this mean?19:35
bswartzshould we assume something has changed w.r.t. networking?19:35
vponomaryovneed to dig up culprit commit19:35
bswartzhave you determined if the breaking change was from neutron or nova?19:35
vponomaryovhaven't looked yet19:35
vponomaryovI suspect Nova19:36
vponomaryovthat could mess up voluem device paths19:36
bswartzyes19:36
bswartzthe nova->cinder attach API changes could be causing it19:37
bswartzit could even be a cinder change causing the problem19:37
vponomaryovfrom logic point of view, neutron is out of suspicion19:38
*** senk has joined #openstack-manila19:38
vponomaryovsince old approach does not work too19:38
*** jprovazn has quit IRC19:39
bswartzit should be pretty easy to reproduce a block device attachment problem if that's the real issue19:39
vponomaryovyes, in case of local testing, I need just to pull latest code and update running services19:40
vponomaryovbut it will be tomorrow19:40
bswartzI was going to suggest a simple devstack install w/ cinder+nova where we replicate the steps that the generic driver does for creating a new volume and attaching it19:41
bswartzmight be simpler to actually use generic driver to do it, but it could be hard to detect when/where it fails19:41
vponomaryovculprit commit was merged between 25th and 29th of November19:44
bswartzvponomaryov: not much happened in cinder during that time window19:45
vponomaryovyes, nothing suspicios from first look19:45
*** catintheroof has quit IRC19:54
openstackgerritVictoria Martinez de la Cruz proposed openstack/manila: Decouple Manila UI from Manila Devstack plugin  https://review.openstack.org/38885519:58
*** gouthamr has joined #openstack-manila20:09
*** dsariel has joined #openstack-manila20:11
*** lpetrut has joined #openstack-manila20:14
*** kaisers_ has joined #openstack-manila20:22
*** porrua has quit IRC20:25
*** senk has quit IRC20:25
*** kaisers_ has quit IRC20:29
*** kaisers_ has joined #openstack-manila20:38
openstackgerritValeriy Ponomaryov proposed openstack/manila: [DNM] test 11  https://review.openstack.org/40496120:46
openstackgerritValeriy Ponomaryov proposed openstack/manila: [DNM] test 12  https://review.openstack.org/40496320:47
openstackgerritValeriy Ponomaryov proposed openstack/manila: [DNM] test 13  https://review.openstack.org/40496420:48
*** catintheroof has joined #openstack-manila20:54
*** xyang_ has joined #openstack-manila21:07
*** alyson_ has quit IRC21:08
*** Yogi1 has quit IRC21:26
openstackgerritGoutham Pacha Ravi proposed openstack/manila-specs: Add a spec to fix and improve Access Rules  https://review.openstack.org/39904921:28
*** dustins has quit IRC21:40
*** gouthamr has quit IRC21:49
*** cknight has quit IRC21:51
*** kaisers_ has quit IRC21:52
*** adrianofr_ has quit IRC21:53
*** kaisers_ has joined #openstack-manila21:54
*** tpsilva has quit IRC21:55
*** vponomaryov1 has joined #openstack-manila21:58
vponomaryov1bswartz: nova broke generic driver, here is proof change: https://review.openstack.org/#/c/404961/ where our job passed reverting nova to the state it had on 24th of November21:59
*** lpetrut has quit IRC22:01
openstackgerritValeriy Ponomaryov proposed openstack/manila: [DNM] test 12  https://review.openstack.org/40496322:04
tbarronvponomaryov1 is a binary search dynamo22:05
*** ganso has quit IRC22:06
openstackgerritValeriy Ponomaryov proposed openstack/manila: [DNM] test 13  https://review.openstack.org/40496422:07
openstackgerritValeriy Ponomaryov proposed openstack/manila: [DNM] test 11  https://review.openstack.org/40496122:07
vponomaryov1tbarron: we do not have time, existing bugs can call some other bugs to their party ))22:12
tbarronvponomaryov1: :)22:12
vponomaryov1tbarron: you are smiling, but it looks like serious Nova bug, where some of their own use cases are broken, that are just not tested in CI22:16
vponomaryov1"just not tested" is control part22:16
tbarronvponomaryov1: might as well smile, but nice work!22:17
tbarronvponomaryov1: do we know *what* use case is broken yet?22:17
vponomaryov1tbarron: assumption: nova VM that has floating IP assigned and one tries to attach more than one volume22:18
vponomaryov1and mount attached volume22:18
vponomaryov1I very very doubt that mount of Cinder volume to filesystem is tested in voting jobs of Nova22:19
vponomaryov1or CInder22:19
tbarronvponomaryov1: do you know yet if it's the attach succeeds or fails?22:19
vponomaryov1attach succeeds22:19
vponomaryov1our logs say that we "hang" on step where we mount volume to filesystem tree22:19
vponomaryov1Hm, I guess, floating IP part can be skipped22:20
vponomaryov1it is just Nova + Cinder22:20
tbarronso it's the mount in the nova instance of a cinder volume that successfully attached, and the second one of these22:21
* tbarron is emulating a rubber duck22:22
vponomaryov1just presence of two attached volumes is enough, I guess22:22
tbarronsome say that's a step *up* the evolutionary chain for him22:23
tbarronI wonder if it has to do with the way cinder "fixed" that two-volume attach issue from a couple weeks ago22:24
vponomaryov1I don't think so22:24
vponomaryov1we know for sure already, that some of "nova changes" is culprit22:25
tbarronk22:25
vponomaryov1but what is strange, that nova's change does not affect DHSS=True mode22:27
vponomaryov1described use case is the same there22:27
vponomaryov1oh no, scenario tests are broken now too -> http://logs.openstack.org/63/404963/2/check/gate-manila-tempest-dsvm-scenario-ubuntu-xenial-nv/e0afac0/logs/testr_results.html.gz  >_<22:36
*** mnaser has quit IRC22:41
*** mnaser has joined #openstack-manila22:49
*** vponomaryov1 has left #openstack-manila23:06
*** StraubTW has quit IRC23:11
*** catintheroof has quit IRC23:12
openstackgerritValeriy Ponomaryov proposed openstack/manila: [DNM] test 11  https://review.openstack.org/40496123:21
*** chlong has quit IRC23:22
openstackgerritValeriy Ponomaryov proposed openstack/manila: [DNM] test 12  https://review.openstack.org/40496323:24
*** eharney has quit IRC23:25
*** xyang_ has quit IRC23:32
*** xyang1 has quit IRC23:50

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