16:02:12 #startmeeting ironic_bfv 16:02:12 Meeting started Thu Apr 20 16:02:12 2017 UTC and is due to finish in 60 minutes. The chair is joanna. Information about MeetBot at http://wiki.debian.org/MeetBot. 16:02:13 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 16:02:16 The meeting name has been set to 'ironic_bfv' 16:02:41 morning! I am sorry for the late start :) 16:03:07 o/ 16:03:17 o/ 16:03:22 #topic Announcements/Reminders 16:04:12 Does anyone have any announcements? 16:04:50 I updated the cinder driver patch yesterday, but would like to discuss the comments later on 16:05:01 There are a couple I didn't address 16:05:28 mjturek: great! I see that in the agenda - will it be fine to discuss in Discussion part? 16:05:36 joanna: totally 16:05:41 great :) 16:05:43 moving on 16:05:45 #topic Current Status 16:05:58 follow up patch for common code got merged :) 16:06:20 great! 16:06:35 \o/ 16:08:38 From what I see in Etherpad, that was the only merged patch since last meeting, other patches are WIP and have new review comments 16:09:08 mjturek: thank you for taking care of the driver patch! 16:09:18 joanna: np it's been really interesting 16:09:30 mjturek: awesome! :) 16:09:41 shall we move to Planning? 16:09:48 +1 16:10:02 #topic Planning/Priorities 16:10:03 shall we update other patches to solve conflicts? 16:10:34 I think that driver patch should stay a priority since it's a direct dependency of the merged patches 16:11:03 agreed 16:11:05 We can think about additional patch to put to weekly pirority list on Monday to accelerate 16:12:22 how about https://review.openstack.org/#/c/406290 ? 16:12:58 joanna: I think there's going to need to be a db patch for getting volume target from volume id (I'd like to handle this). It'll probably be needed to get the driver patch merged 16:13:17 but I'm fine with 406290 16:14:02 i agree, let's rebase and add it to list 16:14:07 mjturek: If you plan to update the wiring patch, you can rebase it on top of your DB reading change, so it won't get lost 16:14:32 surely 16:15:05 mjturek: when the read volume target is ready it can be added to priority list instead. The point is to get reviews as fast as possible :) 16:15:28 mjturek, hshiina: great! 16:15:30 ahhh understood joanna 16:15:53 #info Add https://review.openstack.org/#/c/406290 to weekly priority list 16:16:43 should we move to Discussion? 16:17:14 sure! 16:17:16 yes 16:17:18 #topic Discussion 16:17:36 joanna: hshiina: will be referring to comments here https://review.openstack.org/#/c/366197/39/ironic/drivers/modules/storage/cinder.py 16:17:52 mjturek: do you want to do that one by one? 16:17:58 yeah please 16:18:04 so the first one on line 217 16:18:29 I beleive vdrok is right. The function seems to achieve the same thing as detach_volumes 16:19:07 is anyone opposed to removing the function> 16:19:09 ?* 16:19:21 mjturek: it does it in two ways 16:19:30 first it tries to run with no errors allowed 16:20:04 then, if it fails, it logs that aborting volume attached failed on detach (that's something specfifc) 16:20:14 and then tries once again woth allowing errors 16:20:30 is the error reporting that important though? 16:20:45 that we need a new function for it? 16:21:00 I think the idea of retrying with allowed errors was to detect additional issues with the system 16:21:21 hmmm, alright. I'll bring that up with vdrok and see what he thinks 16:21:28 I'll ping you and hshiina as well to weigh in 16:21:37 I can imagine that if there's something wrong with cinder, so all the ops should fail, as an admin, I'd like to know that 16:22:13 and I think unsuccessful attach may happen even when there are no issues, but if also detach fails that's something worth looking at 16:22:13 yeah I might see the value there 16:23:00 right 16:23:04 but vdrok is right saying that it uses the functionality of detach, it's just a failure handler to add context to logs 16:23:24 got it 16:23:42 we can be less informative, or use these contents inline 16:24:06 However it would be also worth for other possible usages of this function 16:24:27 right right 16:24:39 does it help at all :)? 16:24:57 yeah definitely, but I think I'll need to bring it up with vdrok 16:25:06 before making a decision 16:25:06 sure :) 16:25:10 :) 16:25:21 so moving on to generate_connector comment on 376 16:25:36 ln 376? 16:25:43 yep! 16:26:04 mariojv was asking if Cinder does this. I haven't found anything but haven't looked too much 16:26:33 and I genuinely don't know the answer to 2 or 3 :) 16:26:39 I would assume no to 2 16:26:47 hmm 16:26:52 but was wondering if either of you had insight 16:26:56 if not I'll bug Julia :) 16:27:31 target_iqn should be in volume_target 16:27:59 nova passes ip and iqn to cinder 16:28:51 so she's saying that we need to translate from the connector_id to the IQN 16:29:01 the connector_Id is the IP 16:30:05 but by the time we hit nova we should have the IQN? 16:31:18 and looking at the data structure that this returns, it looks like we'll need IQN 16:31:31 I'm probably going to see if Julia has any insight 16:32:09 mjturek: that is a good idea 16:32:17 cool cool 16:32:22 mjturek: from how I see it we have to pass all the connection info to nova 16:32:30 ahhh 16:32:40 okay so then it is probably a TODO that must be handled 16:32:46 however, it should already be in volume_connector 16:33:01 oh 16:33:14 okay cool 16:33:19 and the IQN can be generated from the other data contained in volume_target, so even if for any reason it's not available, it can be generated 16:33:40 makes sense 16:33:41 (although it might be a good idea to inform that the volume_connector info is incomplete) 16:33:54 s/volume_target/volume_connector/, sorry :) 16:34:07 right :) 16:34:20 okay, makes perfect sense then 16:34:46 so final comment is 409 16:35:29 I think what vdrok means here is that we may have more than one volume connector but one might be broken 16:36:27 would checking len(data) be helpful here? 16:36:32 I see, because vailid variable is reused in the switch case 16:37:15 ahhhh right 16:37:23 yes I think it's a good idea 16:37:23 :) 16:37:34 cool, will do then :) 16:37:56 that's all I had 16:38:03 great! 16:38:14 regarding last issue, i don't think multi volume connectors mean multipath 16:38:57 hshiina: this is about that there might be both ip and iqn for the same volume? 16:39:59 hshiina: please take a look at comment in revision 34 16:40:09 joanna, yes 16:40:17 ln 409 16:42:48 hshiina: does it answer the quesion? 16:45:14 joanna, i haven't fully understood multipath. but, multipath seems more complicated 16:46:05 so it sounds like it simply means that multiple paths are available to the volume. The fact that an IP and IQN are available means multiple paths are available, even if it's just for one driver 16:46:11 one volume* 16:46:19 from what I understand what Julia is saying, there is no harm in setting multipath to true, as driver should handle that 16:48:12 mjturek: I think that since it's confusing, maybe it's a goos idea to add a comment there explaining why multipath is handled this way? 16:48:43 also, can you confirm with Julia that it's a proper way to handle multipath? 16:49:26 joanna: agreed 16:49:39 hshiina: are you ok with that? :) 16:50:29 multipath is listed in feature capabilities in the approved spce: https://specs.openstack.org/openstack/ironic-specs/specs/not-implemented/boot-from-volume-reference-drivers.html 16:50:58 joanna, it's ok to confirm with Julia 16:52:05 hshiina: it's listed here as a potential capability - so maybe it's a good idea to have a happy scenario now, and when it's ready add better support for multipath if what we have is insufficient? 16:53:28 joanna, yes 16:53:58 mjturek: if that will be fine with Julia, this comment may be a TODO comment for the future, then :) 16:54:14 cool! that was the only item, should we move to open discussion? 16:54:24 sure, I'll reach out to her later and ping hshiina as well 16:54:32 mjturek: that's awesome! 16:54:32 +1 16:54:41 mjturek, thanks 16:54:47 #info mjturek to follow up on multipath with hshiina and TheJulia 16:54:51 #topic Open Discussion 16:55:44 so this might be the last BFV meeting I attend. I will miss working on it! Please do not hesitate to contact me with any questions or anything :) 16:56:22 joanna: :( will do 16:56:37 joanna: do we need someone to facilitate the meeting> 16:56:47 I'd be happy to help unless hshiina would rather do it 16:57:15 TheJulia should be back next week 16:57:22 so I think there's no such need :) 16:57:26 joanna: okay great! 16:57:31 awesome! 16:57:34 so are we done? 16:57:46 yes. 16:57:51 great! 16:57:54 ttyl all 16:57:57 good luck & have fun! :) 16:58:04 joanna, thank you for your work! 16:58:04 you too joanna :) 16:58:04 thank you :) 16:58:05 #endmeeting