16:01:03 <TheJulia> #startmeeting ironic_bfv
16:01:05 <openstack> Meeting started Thu May  4 16:01:03 2017 UTC and is due to finish in 60 minutes.  The chair is TheJulia. Information about MeetBot at http://wiki.debian.org/MeetBot.
16:01:06 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
16:01:08 <openstack> The meeting name has been set to 'ironic_bfv'
16:01:33 <hshiina> o/
16:01:39 <derekh> o/
16:01:42 <TheJulia> As always our agenda is at https://wiki.openstack.org/wiki/Meetings/Ironic-BFV
16:01:46 <TheJulia> #link https://wiki.openstack.org/wiki/Meetings/Ironic-BFV
16:02:11 <TheJulia> I forgot to clean it from our last meeting, but it looks like we have one additional item that mjturek put on it since last week.
16:02:30 <TheJulia> #topic Announcements/Reminders
16:02:31 <wanyen> o/
16:03:03 <TheJulia> Next week is the Summit, as all of the ironic meetings are being cancelled as a number of participants will be at the summit or otherwise occupied next week.
16:03:11 <TheJulia> #info Meeting next week is cancelled.
16:03:15 <TheJulia> #undo
16:03:16 <openstack> Removing item from minutes: #info Meeting next week is cancelled.
16:03:21 <TheJulia> #info Meeting next week is cancelled due to the Summit.
16:03:37 <TheJulia> Anyone else have any announcements?
16:05:20 <TheJulia> I guess not, so moving on...
16:05:27 <mjturek> o/
16:05:47 <TheJulia> \o mjturek
16:05:58 <TheJulia> #topic Current Status
16:06:10 <TheJulia> #link https://etherpad.openstack.org/p/Ironic-BFV
16:06:54 <mjturek> So I submitted an update to the cinder driver. I didn't handle dtantsur's comments yet but will. Let me know what you think of how I handled the wwpn/wwnn requirement
16:06:59 <mjturek> not sure it's enough
16:07:20 <TheJulia> mjturek: Thank you!  I saw it posted but I've not had a chance to look
16:07:30 <mjturek> cool cool :)
16:07:59 <TheJulia> I rebased the storage interface wire-in and the pxe template write-out changes for ipxe based boot
16:08:06 <dtantsur> o/
16:08:13 <TheJulia> \o dtantsur
16:08:56 <stendulker> Julia, there is a ironic spec patch related to UEFI iscsi boot https://review.openstack.org/207337
16:09:07 <stendulker> Can it be added toetherpad?
16:09:24 <stendulker> Its a generic spec for UEFI iscsi boot
16:10:12 <TheJulia> stendulker: Sure, but under a vendor implementation section
16:10:41 <TheJulia> mjturek: you mention wwpn/wwnn requirement, can you elaborate or is that in the comments on the cinder storage interface driver?
16:10:43 <stendulker> TheJulia: sure. Will do
16:10:49 <stendulker> TheJulia: Thanks
16:10:57 <mjturek> I can elaborate
16:11:21 <mjturek> so if you're out of the loop, cinder requires both wwpn and wwnn for a valid FCoE connector
16:11:50 <TheJulia> (or FC) :)
16:12:05 <mjturek> the way I handled that in ironic's cinder driver is by counting how many wwpns and wwnns we have. If their equal, I consider it valid
16:12:28 <mjturek> but I'm not sure that's enough
16:12:56 <wanyen> The Julia, The spec https://review.openstack.org/207337 is generic not vendor specific but the implementation can be vendor specific.
16:12:58 <TheJulia> That is a good question
16:12:59 <mjturek> (mostly because I'm not super familiar with FC :-\)
16:14:03 <TheJulia> I _think_ that might be okay, we would likely just want to document that is the case
16:14:12 <TheJulia> and we can always revise it later if we must
16:14:30 <mjturek> cool, I think so as well, just wasn't 100%
16:14:50 <mjturek> it might be more obvious once everything's wired together properly
16:15:29 <TheJulia> wanyen: With it on the list, It should get some reviews moving forward.  I'll try to take some time in the next week to take a look at it.
16:15:46 <wanyen> The Julia, ty
16:15:49 <TheJulia> So, I think that covers current status
16:15:57 <TheJulia> Anyone have anything else status wise?
16:16:50 <TheJulia> I'm guessing not, so moving on...
16:17:46 <TheJulia> #topic Priorities
16:17:57 <TheJulia> I think the number 1 priority is https://review.openstack.org/#/c/366197/ at this point in time.
16:17:58 <TheJulia> #link https://review.openstack.org/#/c/366197/
16:18:24 <mjturek> +1
16:19:05 <hshiina> +1
16:19:40 <TheJulia> I don't really think there is anything else at the moment on the priorities front. After that would naturally be the interface wire-in.
16:20:36 <TheJulia> So, discussion time unless anyone has anything to add/bring up/inquire about?
16:21:26 <hshiina> is it necessary for cinder driver patch? https://review.openstack.org/#/c/460250/
16:21:48 <mjturek> hshiina: I'm planning on it being necessary
16:21:58 <mjturek> I haven't put it to use yet
16:22:06 <mjturek> I'll try to get that in by EoD
16:22:15 <TheJulia> mjturek: Awesome, Thank you
16:22:23 <hshiina> mjturek, thanks
16:22:29 <TheJulia> #topic Discussion
16:22:44 <TheJulia> I believe we really only have one topic, raised by mjturek for Multipath.
16:22:53 * TheJulia hands mjturek the microphone
16:23:09 <mjturek> yep, so we had a discussion about multipath a couple weeks back
16:23:41 <mjturek> vdrok had suggested something - let me grab it
16:24:00 <mjturek> https://review.openstack.org/#/c/366197/38/ironic/drivers/modules/storage/cinder.py see his comment on 414
16:24:15 <mjturek> "But what if only one of them is valid? Maybe we should be checking len(data) here?"
16:25:06 <mjturek> this led to another question from hshiina
16:25:38 <TheJulia> Ahh, right
16:25:52 <mjturek> which was, if I remember correctly "does multiple connectors really mean multiple paths?"
16:26:12 <TheJulia> so it would end up needing to be >1 iqn and or >1 wwpns
16:26:39 <mjturek> that makes sense
16:26:39 <TheJulia> I think... for iscsi it would likely require double checking RFCs
16:27:03 <TheJulia> we should appropriately set the flag for whatever cinder driver is consuming the data though
16:27:34 <mjturek> hmmm - would we know what driver we were sending to?
16:28:27 <TheJulia> afaik, there is absolutely no way to know
16:28:46 <mjturek> ah
16:29:52 <TheJulia> I seem to remember the field that might indicate it could actually be anything
16:29:59 <mjturek> lolol
16:30:07 <TheJulia> as it ends up becoming user defined
16:30:13 <TheJulia> err, operator defined
16:30:32 <mjturek> okay - so I guess need to look at how cinder leverages the field
16:30:46 <TheJulia> I seem to remember inline notes in the code to that effect
16:31:00 <mjturek> cool, I'll take a look
16:31:10 <TheJulia> but you'll spot it quickly when you look at the volume fields
16:31:27 <mjturek> but to us, multipath means that FC and iSCSI paths are available?
16:31:33 <mjturek> not even necessarily to the same volume?
16:32:06 <TheJulia> multipath is that the node can support it, so if it has multiple valid means, then we should indicate that and all of the pertinent data to support it
16:32:40 <mjturek> got it.
16:32:50 <TheJulia> okay
16:33:02 <TheJulia> Anything else for multipath?
16:33:51 <mjturek> I'm good for now
16:34:31 <TheJulia> Awesome, well, if there are no objections, then we will go to open discussion, and close out the meeting if there is nothing to discuss then.
16:35:22 <TheJulia> #topic Open Discussion
16:36:48 * TheJulia gets out the appropriate time ticking away theme song :)
16:38:26 <TheJulia> If there is nothing else, everyone can get 20 minutes back. :)
16:39:05 <mjturek> lol, I'm good :)
16:39:35 * dtantsur too
16:39:53 <TheJulia> Awesome, have a wonderful week everyone!
16:39:58 <TheJulia> #endmeeting