| *** lukebrowning has quit IRC | 00:01 | |
| *** lukebrowning has joined #openstack-tc | 00:02 | |
| *** kumarmn has joined #openstack-tc | 00:07 | |
| *** lukebrowning has quit IRC | 00:07 | |
| *** lukebrowning has joined #openstack-tc | 00:19 | |
| *** lukebrowning has quit IRC | 00:24 | |
| *** lukebrowning has joined #openstack-tc | 00:26 | |
| *** lukebrowning has quit IRC | 00:30 | |
| *** lukebrowning has joined #openstack-tc | 00:32 | |
| *** lukebrowning has quit IRC | 00:37 | |
| *** lukebrowning has joined #openstack-tc | 00:38 | |
| *** kumarmn has quit IRC | 00:41 | |
| *** lukebrowning has quit IRC | 00:43 | |
| *** lukebrowning has joined #openstack-tc | 00:45 | |
| *** lukebrowning has quit IRC | 00:49 | |
| *** lukebrowning has joined #openstack-tc | 00:51 | |
| *** lukebrowning has quit IRC | 00:56 | |
| *** lukebrowning has joined #openstack-tc | 00:57 | |
| *** lukebrowning has quit IRC | 01:01 | |
| fungi | tc-members: commence to be doing officey hour things if you like | 01:02 |
|---|---|---|
| fungi | though we talked the heck out of some stuff earlier, so i can understand if that's covered | 01:03 |
| *** lukebrowning has joined #openstack-tc | 01:03 | |
| *** lukebrowning has quit IRC | 01:08 | |
| *** lukebrowning has joined #openstack-tc | 01:19 | |
| *** lukebrowning has quit IRC | 01:24 | |
| *** lukebrowning has joined #openstack-tc | 01:26 | |
| *** liujiong has joined #openstack-tc | 01:29 | |
| *** lukebrowning has quit IRC | 01:30 | |
| *** lukebrowning has joined #openstack-tc | 01:32 | |
| *** lukebrowning has quit IRC | 01:37 | |
| *** lukebrowning has joined #openstack-tc | 01:38 | |
| *** lukebrowning has quit IRC | 01:43 | |
| *** lukebrowning has joined #openstack-tc | 01:44 | |
| *** lukebrowning has quit IRC | 01:49 | |
| *** lukebrowning has joined #openstack-tc | 01:51 | |
| *** lukebrowning has quit IRC | 01:55 | |
| *** liujiong has quit IRC | 01:56 | |
| *** liujiong has joined #openstack-tc | 01:56 | |
| *** lukebrowning has joined #openstack-tc | 01:57 | |
| *** edleafe has quit IRC | 01:57 | |
| fungi | well now, that was simply riveting! | 02:00 |
| fungi | join us thursday at 15:00 utc for our third and final office hour of the week | 02:01 |
| *** lukebrowning has quit IRC | 02:02 | |
| *** lukebrowning has joined #openstack-tc | 02:04 | |
| *** lukebrowning has quit IRC | 02:08 | |
| *** lukebrowning has joined #openstack-tc | 02:19 | |
| *** lukebrowning has quit IRC | 02:24 | |
| *** lukebrowning has joined #openstack-tc | 02:25 | |
| *** lukebrowning has quit IRC | 02:30 | |
| *** lukebrowning has joined #openstack-tc | 02:32 | |
| *** lukebrowning has quit IRC | 02:36 | |
| *** lukebrowning has joined #openstack-tc | 02:44 | |
| *** lukebrowning has quit IRC | 02:49 | |
| *** lukebrowning has joined #openstack-tc | 02:50 | |
| *** lukebrowning has quit IRC | 02:55 | |
| *** lukebrowning has joined #openstack-tc | 03:01 | |
| *** lukebrowning has quit IRC | 03:06 | |
| *** lukebrowning has joined #openstack-tc | 03:07 | |
| *** lukebrowning has quit IRC | 03:12 | |
| *** lukebrowning has joined #openstack-tc | 03:14 | |
| *** lukebrowning has quit IRC | 03:18 | |
| *** lukebrowning has joined #openstack-tc | 03:20 | |
| *** lukebrowning has quit IRC | 03:25 | |
| *** lukebrowning has joined #openstack-tc | 03:26 | |
| *** lukebrowning has quit IRC | 03:31 | |
| *** lukebrowning has joined #openstack-tc | 03:33 | |
| *** lukebrowning has quit IRC | 03:37 | |
| *** lukebrowning has joined #openstack-tc | 04:02 | |
| *** lukebrowning has quit IRC | 04:07 | |
| *** lukebrowning has joined #openstack-tc | 04:13 | |
| *** lukebrowning has quit IRC | 04:18 | |
| *** lukebrowning has joined #openstack-tc | 04:19 | |
| *** lukebrowning has quit IRC | 04:24 | |
| *** lukebrowning has joined #openstack-tc | 04:25 | |
| *** lukebrowning has quit IRC | 04:30 | |
| *** lukebrowning has joined #openstack-tc | 04:32 | |
| *** kumarmn has joined #openstack-tc | 04:33 | |
| *** lukebrowning has quit IRC | 04:36 | |
| *** lukebrowning has joined #openstack-tc | 04:38 | |
| *** lukebrowning has quit IRC | 04:42 | |
| *** lukebrowning has joined #openstack-tc | 04:50 | |
| *** lukebrowning has quit IRC | 04:55 | |
| *** lukebrowning has joined #openstack-tc | 04:57 | |
| *** ianw is now known as ianw_pto | 04:57 | |
| *** kumarmn has quit IRC | 05:02 | |
| *** lukebrowning has quit IRC | 05:06 | |
| *** lukebrowning has joined #openstack-tc | 05:08 | |
| *** lukebrowning has quit IRC | 05:13 | |
| *** lukebrowning has joined #openstack-tc | 05:14 | |
| *** lukebrowning has quit IRC | 05:19 | |
| *** lukebrowning has joined #openstack-tc | 05:21 | |
| *** lukebrowning has quit IRC | 05:25 | |
| *** lukebrowning has joined #openstack-tc | 05:27 | |
| *** lukebrowning has quit IRC | 05:32 | |
| *** lukebrowning has joined #openstack-tc | 05:39 | |
| *** lukebrowning has quit IRC | 05:44 | |
| *** lukebrowning has joined #openstack-tc | 05:46 | |
| *** kumarmn has joined #openstack-tc | 05:50 | |
| *** lukebrowning has quit IRC | 05:50 | |
| *** lukebrowning has joined #openstack-tc | 05:52 | |
| *** kumarmn has quit IRC | 05:55 | |
| *** lukebrowning has quit IRC | 05:57 | |
| *** lukebrowning has joined #openstack-tc | 05:58 | |
| *** lukebrowning has quit IRC | 06:03 | |
| *** lukebrowning has joined #openstack-tc | 06:05 | |
| *** lukebrowning has quit IRC | 06:09 | |
| *** lukebrowning has joined #openstack-tc | 06:14 | |
| *** lukebrowning has quit IRC | 06:18 | |
| *** lukebrowning has joined #openstack-tc | 06:25 | |
| *** lukebrowning has quit IRC | 06:30 | |
| *** lukebrowning has joined #openstack-tc | 06:31 | |
| *** lukebrowning has quit IRC | 06:41 | |
| *** ianychoi has quit IRC | 06:44 | |
| *** lukebrowning has joined #openstack-tc | 06:45 | |
| *** lukebrowning has quit IRC | 06:50 | |
| *** kumarmn has joined #openstack-tc | 06:51 | |
| *** lukebrowning has joined #openstack-tc | 06:52 | |
| *** kumarmn has quit IRC | 06:55 | |
| *** lukebrowning has quit IRC | 06:57 | |
| *** lukebrowning has joined #openstack-tc | 06:58 | |
| *** lukebrowning has quit IRC | 07:03 | |
| *** lukebrowning has joined #openstack-tc | 07:04 | |
| *** lukebrowning has quit IRC | 07:09 | |
| *** lukebrowning has joined #openstack-tc | 07:13 | |
| *** lukebrowning has quit IRC | 07:18 | |
| *** lukebrowning has joined #openstack-tc | 07:20 | |
| *** lukebrowning has quit IRC | 07:24 | |
| *** lukebrowning has joined #openstack-tc | 07:26 | |
| *** lukebrowning has quit IRC | 07:31 | |
| *** lukebrowning has joined #openstack-tc | 07:32 | |
| *** lukebrowning has quit IRC | 07:37 | |
| *** lukebrowning has joined #openstack-tc | 07:38 | |
| *** lukebrowning has quit IRC | 07:43 | |
| *** lukebrowning has joined #openstack-tc | 07:45 | |
| *** lukebrowning has quit IRC | 07:50 | |
| *** lukebrowning has joined #openstack-tc | 07:51 | |
| *** kumarmn has joined #openstack-tc | 07:51 | |
| *** lukebrowning has quit IRC | 07:56 | |
| *** kumarmn has quit IRC | 07:57 | |
| *** lukebrowning has joined #openstack-tc | 07:57 | |
| *** lukebrowning has quit IRC | 08:02 | |
| *** lukebrowning has joined #openstack-tc | 08:04 | |
| *** lukebrowning has quit IRC | 08:09 | |
| *** lukebrowning has joined #openstack-tc | 08:16 | |
| *** lukebrowning has quit IRC | 08:21 | |
| *** lukebrowning has joined #openstack-tc | 08:23 | |
| *** lukebrowning has quit IRC | 08:27 | |
| *** lukebrowning has joined #openstack-tc | 08:29 | |
| *** lukebrowning has quit IRC | 08:34 | |
| *** lukebrowning has joined #openstack-tc | 08:35 | |
| *** lukebrowning has quit IRC | 08:40 | |
| *** kmalloc has quit IRC | 08:41 | |
| *** lukebrowning has joined #openstack-tc | 08:42 | |
| *** lukebrowning has quit IRC | 08:46 | |
| *** lukebrowning has joined #openstack-tc | 08:48 | |
| *** kumarmn has joined #openstack-tc | 08:52 | |
| *** lukebrowning has quit IRC | 08:53 | |
| *** lukebrowning has joined #openstack-tc | 08:54 | |
| *** kumarmn has quit IRC | 08:57 | |
| *** lukebrowning has quit IRC | 08:59 | |
| *** lukebrowning has joined #openstack-tc | 09:00 | |
| *** lukebrowning has quit IRC | 09:05 | |
| *** lukebrowning has joined #openstack-tc | 09:07 | |
| *** lukebrowning has quit IRC | 09:12 | |
| *** lukebrowning has joined #openstack-tc | 09:19 | |
| *** lukebrowning has quit IRC | 09:24 | |
| *** lukebrowning has joined #openstack-tc | 09:30 | |
| *** lukebrowning has quit IRC | 09:34 | |
| *** lukebrowning has joined #openstack-tc | 09:36 | |
| *** lukebrowning has quit IRC | 09:41 | |
| *** lukebrowning has joined #openstack-tc | 09:42 | |
| *** lukebrowning has quit IRC | 09:47 | |
| *** lukebrowning has joined #openstack-tc | 09:49 | |
| *** lukebrowning has quit IRC | 09:53 | |
| *** lukebrowning has joined #openstack-tc | 09:55 | |
| *** lukebrowning has quit IRC | 10:00 | |
| *** lukebrowning has joined #openstack-tc | 10:01 | |
| *** lukebrowning has quit IRC | 10:06 | |
| *** liujiong has quit IRC | 10:06 | |
| *** lukebrowning has joined #openstack-tc | 10:24 | |
| *** lukebrowning has quit IRC | 10:29 | |
| *** kumarmn has joined #openstack-tc | 10:54 | |
| *** kumarmn has quit IRC | 10:59 | |
| *** dtantsur|afk is now known as dtantsur | 11:02 | |
| *** lukebrowning has joined #openstack-tc | 11:05 | |
| *** lukebrowning has quit IRC | 11:10 | |
| *** gcb has quit IRC | 11:12 | |
| *** lukebrowning has joined #openstack-tc | 11:16 | |
| *** lukebrowning has quit IRC | 11:21 | |
| *** lukebrowning has joined #openstack-tc | 11:23 | |
| *** lukebrowning has quit IRC | 11:27 | |
| *** lukebrowning has joined #openstack-tc | 11:29 | |
| *** lukebrowning has quit IRC | 11:34 | |
| *** lukebrowning has joined #openstack-tc | 11:35 | |
| *** lukebrowning has quit IRC | 11:40 | |
| *** lukebrowning has joined #openstack-tc | 11:41 | |
| *** lukebrowning has quit IRC | 11:46 | |
| *** lukebrowning has joined #openstack-tc | 11:48 | |
| *** lukebrowning has quit IRC | 11:53 | |
| *** lukebrowning has joined #openstack-tc | 11:54 | |
| *** kumarmn has joined #openstack-tc | 11:55 | |
| *** cdent has joined #openstack-tc | 11:57 | |
| *** lukebrowning has quit IRC | 11:59 | |
| *** kumarmn has quit IRC | 12:00 | |
| *** lukebrowning has joined #openstack-tc | 12:00 | |
| *** kumarmn has joined #openstack-tc | 12:02 | |
| *** lukebrowning has quit IRC | 12:05 | |
| *** lukebrowning has joined #openstack-tc | 12:07 | |
| *** kumarmn has quit IRC | 12:07 | |
| *** lukebrowning has quit IRC | 12:11 | |
| dims | fungi :) | 12:12 |
| cdent | [t 1I4l] | 12:16 |
| purplerbot | <fungi> well now, that was simply riveting! [2017-12-20 02:00:18.070984] [n 1I4l] | 12:16 |
| *** gcb has joined #openstack-tc | 12:39 | |
| *** lukebrowning has joined #openstack-tc | 12:45 | |
| *** gcb has quit IRC | 12:47 | |
| *** lukebrowning has quit IRC | 12:49 | |
| *** lukebrowning has joined #openstack-tc | 12:55 | |
| *** lukebrowning has quit IRC | 13:00 | |
| *** lukebrowning has joined #openstack-tc | 13:01 | |
| *** kumarmn has joined #openstack-tc | 13:03 | |
| *** lukebrowning has quit IRC | 13:06 | |
| *** lukebrowning has joined #openstack-tc | 13:07 | |
| *** kumarmn has quit IRC | 13:08 | |
| *** lukebrowning has quit IRC | 13:12 | |
| *** openstackgerrit has quit IRC | 13:13 | |
| -openstackstatus- NOTICE: gerrit is being restarted due to extreme slowness | 13:15 | |
| *** lukebrowning has joined #openstack-tc | 13:18 | |
| *** lukebrowning has quit IRC | 13:23 | |
| *** lukebrowning has joined #openstack-tc | 13:25 | |
| *** lukebrowning_ has joined #openstack-tc | 13:28 | |
| *** lukebrowning has quit IRC | 13:29 | |
| *** lukebrowning_ has quit IRC | 13:33 | |
| *** lukebrowning has joined #openstack-tc | 13:34 | |
| *** lukebrowning has quit IRC | 13:39 | |
| *** lukebrowning has joined #openstack-tc | 13:40 | |
| *** lukebrowning has quit IRC | 13:45 | |
| *** lukebrowning has joined #openstack-tc | 13:46 | |
| *** cdent has quit IRC | 13:47 | |
| *** kumarmn has joined #openstack-tc | 13:49 | |
| *** lukebrowning has quit IRC | 13:51 | |
| *** mriedem1 has joined #openstack-tc | 13:52 | |
| *** cdent has joined #openstack-tc | 13:53 | |
| *** lukebrowning has joined #openstack-tc | 13:53 | |
| *** kumarmn has quit IRC | 13:54 | |
| *** mriedem has quit IRC | 13:55 | |
| *** mriedem1 is now known as mriedem | 13:56 | |
| *** lukebrowning has quit IRC | 13:58 | |
| *** lukebrowning has joined #openstack-tc | 14:00 | |
| *** lukebrowning has quit IRC | 14:05 | |
| *** lukebrowning has joined #openstack-tc | 14:06 | |
| *** lukebrowning has quit IRC | 14:11 | |
| *** lukebrowning has joined #openstack-tc | 14:12 | |
| *** lukebrowning has quit IRC | 14:17 | |
| *** lukebrowning has joined #openstack-tc | 14:19 | |
| *** lukebrowning has quit IRC | 14:24 | |
| *** lukebrowning has joined #openstack-tc | 14:25 | |
| *** lukebrowning has quit IRC | 14:30 | |
| *** lukebrowning has joined #openstack-tc | 14:31 | |
| smcginnis | It's not always that you get the whole office to yourself. | 14:35 |
| *** lukebrowning has quit IRC | 14:36 | |
| *** lukebrowning has joined #openstack-tc | 14:38 | |
| *** lukebrowning has quit IRC | 14:42 | |
| *** kumarmn has joined #openstack-tc | 14:44 | |
| TheJulia | Except then all work can be assigned to those not there <.< >.> | 14:44 |
| * cdent always gets the whole office | 14:46 | |
| cdent | this is a combo of great and tragic | 14:46 |
| fungi | TheJulia: luckily, all attendees were smart enough not to assign me anything ;) | 14:46 |
| fungi | oh, i see, you're saying i should have taken the opportunity to "delegate" | 14:47 |
| *** lukebrowning has joined #openstack-tc | 14:50 | |
| TheJulia | lol | 14:55 |
| *** lukebrowning has quit IRC | 14:55 | |
| *** lukebrowning has joined #openstack-tc | 14:57 | |
| *** mriedem is now known as mriedem_away | 14:57 | |
| *** lukebrowning has quit IRC | 15:01 | |
| *** lukebrowning has joined #openstack-tc | 15:03 | |
| *** marst has joined #openstack-tc | 15:06 | |
| *** lukebrowning has quit IRC | 15:08 | |
| *** lukebrowning has joined #openstack-tc | 15:09 | |
| *** lukebrowning has quit IRC | 15:13 | |
| *** lukebrowning has joined #openstack-tc | 15:41 | |
| *** lukebrowning has quit IRC | 15:45 | |
| *** lukebrowning has joined #openstack-tc | 15:47 | |
| *** lukebrowning has quit IRC | 15:52 | |
| *** lukebrowning has joined #openstack-tc | 15:53 | |
| *** lukebrowning has quit IRC | 15:58 | |
| *** lukebrowning has joined #openstack-tc | 16:00 | |
| *** lukebrowning has quit IRC | 16:04 | |
| *** lukebrowning has joined #openstack-tc | 16:06 | |
| *** mriedem_away is now known as mriedem | 16:09 | |
| *** lukebrowning has quit IRC | 16:10 | |
| *** lukebrowning has joined #openstack-tc | 16:12 | |
| *** lukebrowning has quit IRC | 16:17 | |
| *** lukebrowning has joined #openstack-tc | 16:19 | |
| *** kumarmn has quit IRC | 16:22 | |
| *** lukebrowning has quit IRC | 16:24 | |
| *** lukebrowning has joined #openstack-tc | 16:26 | |
| *** lukebrowning has quit IRC | 16:31 | |
| *** lukebrowning has joined #openstack-tc | 16:32 | |
| *** lukebrowning has quit IRC | 16:37 | |
| *** lukebrowning has joined #openstack-tc | 16:38 | |
| *** kumarmn has joined #openstack-tc | 16:39 | |
| *** lukebrowning has quit IRC | 16:43 | |
| *** lukebrowning has joined #openstack-tc | 16:44 | |
| *** lukebrowning has quit IRC | 16:49 | |
| *** lukebrowning has joined #openstack-tc | 16:51 | |
| *** lukebrowning has quit IRC | 16:55 | |
| *** dtantsur is now known as dtantsur|afk | 17:34 | |
| *** lukebrowning has joined #openstack-tc | 17:40 | |
| *** lukebrowning has quit IRC | 17:45 | |
| *** lukebrowning has joined #openstack-tc | 17:48 | |
| *** lukebrowning has quit IRC | 17:52 | |
| *** lukebrowning has joined #openstack-tc | 17:54 | |
| *** lukebrowning has quit IRC | 17:59 | |
| *** lukebrowning has joined #openstack-tc | 18:00 | |
| *** lukebrowning has quit IRC | 18:05 | |
| *** lukebrowning has joined #openstack-tc | 18:07 | |
| *** lukebrowning has quit IRC | 18:11 | |
| *** lukebrowning has joined #openstack-tc | 18:13 | |
| harlowja | ok, we make all the decision now right? | 18:16 |
| *** lukebrowning has quit IRC | 18:18 | |
| harlowja | *all of them* lol | 18:25 |
| harlowja | while everyone is on vacation | 18:25 |
| cdent | decided: harlowja for Acquisition Editor | 18:26 |
| harlowja | eck | 18:26 |
| harlowja | lol | 18:26 |
| dims | hey harlowja | 18:27 |
| harlowja | hola | 18:29 |
| harlowja | como etas | 18:29 |
| harlowja | *como estas | 18:29 |
| harlowja | so ya, lots of backscroll reading, lol | 18:40 |
| TheJulia | harlowja: I guess we're sufficiently in the quiet time of the year to begin making all of the decisions right now... First decision though, we must laugh evilly! | 18:43 |
| *** lukebrowning has joined #openstack-tc | 18:43 | |
| *** lukebrowning has quit IRC | 18:48 | |
| harlowja | yup | 18:50 |
| harlowja | do it | 18:50 |
| harlowja | so i suggest a marketing campaign done by openstack foundation to showcase who is using openstack and why and <generate some noise> | 18:50 |
| harlowja | that's my first suggestion, lol | 18:50 |
| *** lukebrowning has joined #openstack-tc | 18:51 | |
| harlowja | i've seen to many of https://www.makeway.world/wp-content/uploads/2017/10/build-on-aws-header.png at airports ,lol | 18:51 |
| harlowja | ^ recently | 18:51 |
| harlowja | and https://pbs.twimg.com/media/DKRIc5GW0AArw8L.jpg and ... | 18:51 |
| harlowja | i believe that will help fix a root cause (publicity and coolness effect, call it whatever u want, PR, lol) | 18:52 |
| TheJulia | Heh, So we can't just have the hypnotoad in airports? | 18:53 |
| *** lukebrowning has quit IRC | 18:55 | |
| *** lukebrowning has joined #openstack-tc | 18:57 | |
| harlowja | that'd be fine with me | 19:00 |
| harlowja | lol | 19:00 |
| *** lukebrowning has quit IRC | 19:01 | |
| *** lukebrowning has joined #openstack-tc | 19:03 | |
| persia | Hype about OpenStack doesn't necessarily translate into FTEs assigned to develop OpenStack. It did before, but it would need significant VC (internal or external) to cause a direct mapping, and I think they learned a lesson already (or think they have) | 19:05 |
| harlowja | we need to rebrand then, lol | 19:07 |
| harlowja | StackStackStack | 19:07 |
| *** lukebrowning has quit IRC | 19:08 | |
| *** lukebrowning has joined #openstack-tc | 19:09 | |
| harlowja | do we know that the hype doesn't translate? | 19:12 |
| harlowja | or PR hype or... | 19:13 |
| harlowja | what lesson did said VC (or whoever) learn? | 19:13 |
| harlowja | or what supposed lesson | 19:13 |
| persia | In other environments, press hype usually correlates with folk making investment in an attempt to extract revenues from that which is hyped. For most software, once a working example is present, investment tends to be on the sales side, rather than on the development side. These are only general guidelines: they may not apply in specific circumstances, although the transition we have seen in vendor investment in openstack over the past couple | 19:14 |
| persia | years follows that pattern. | 19:14 |
| *** lukebrowning has quit IRC | 19:14 | |
| persia | I suspect the VCs do not currently believe "OpenStack" provides a significant multiple for investment turnover, and that most current investors are supporting active business processes rather than speculative valuation. I could be mistaken. | 19:15 |
| *** lukebrowning has joined #openstack-tc | 19:15 | |
| mrhillsman | i think we limit most of our "hype" discussions to US when in actuality OpenStack is still very much "hype" for non-US folks | 19:16 |
| harlowja | Ya, intersting, i think this is one of the root causes (call it whatever u want), others may or may not agree | 19:16 |
| harlowja | mrhillsman agreed | 19:16 |
| harlowja | but i'm just in US, ha | 19:16 |
| mrhillsman | we need to do a better job fostering non-US investment and developers | 19:17 |
| mrhillsman | which unfortunately is difficult because of a number of reasons | 19:17 |
| mrhillsman | but is definitely something we should as a community do (top-down/bottom-up) | 19:17 |
| persia | mrhillsman: While OpenStack is very popular in non-US, do you perceive significant VC-style investment? | 19:17 |
| * persia mostly sees non-US firms developing business processes that use OpenStack successfully, rather than many firms trying to resell OpenStack in non-US | 19:18 | |
| mrhillsman | i do not have enough of a view to offer any facts but based on opinion i think it is a mixture of both | 19:19 |
| harlowja | does openstack foundation have a PR person/marketing manager that is responsible for trying to lead/figure out all of this? | 19:19 |
| harlowja | ^ i don't really know, lol | 19:19 |
| mrhillsman | but i still believe that the issue is fostering one or the other | 19:20 |
| *** lukebrowning has quit IRC | 19:20 | |
| mrhillsman | or both for that matter and not something to establish trends or make facts about as a basis for the underlying community/project/etc | 19:21 |
| mrhillsman | i could be saying that wrong | 19:21 |
| harlowja | ya, seems relevant here, lol | 19:22 |
| mrhillsman | but basically i think because openstack and other open source projects for that matter are by default global we do specific projects and open source at large a disservice when we limit our observation through US filter | 19:22 |
| harlowja | sureeeee :) | 19:22 |
| *** lukebrowning has joined #openstack-tc | 19:24 | |
| *** lukebrowning has quit IRC | 19:29 | |
| *** lukebrowning has joined #openstack-tc | 19:30 | |
| *** lukebrowning has quit IRC | 19:35 | |
| harlowja | as for the PTL stuff, i might just wait to give my thoughts on that for later / meeting-hours | 19:36 |
| harlowja | some of this requires some meditation, lol | 19:36 |
| *** lukebrowning has joined #openstack-tc | 19:37 | |
| *** lukebrowning has quit IRC | 19:42 | |
| *** lukebrowning has joined #openstack-tc | 19:43 | |
| harlowja | i'd also be +1 for having all TC office hours with video | 19:47 |
| harlowja | that would make the IRC logs more interesting to read | 19:47 |
| harlowja | so ya, please do that :) | 19:47 |
| harlowja | in the style of https://upload.wikimedia.org/wikipedia/en/5/56/Brady_Bunch.jpg would be nice | 19:48 |
| cdent | but then I'd have to wear pants | 19:48 |
| harlowja | nah nah, just your face | 19:48 |
| harlowja | ^ see above pict, lol | 19:48 |
| cdent | you don't wear pants on your face? | 19:49 |
| harlowja | only on sundays | 19:49 |
| *** lukebrowning has quit IRC | 19:51 | |
| *** lukebrowning has joined #openstack-tc | 19:52 | |
| cdent | I think I'm going to crash out of today | 19:54 |
| harlowja | picts please | 19:54 |
| harlowja | i recommend mandatory gopros for all TC members | 19:54 |
| harlowja | lol | 19:54 |
| harlowja | twitch streams perhaps? | 19:54 |
| harlowja | lol | 19:54 |
| harlowja | maybe that's what we should all do | 19:55 |
| harlowja | every dev should have a twitch stream | 19:55 |
| cdent | no sir | 19:55 |
| harlowja | lol | 19:55 |
| cdent | night night | 19:56 |
| *** cdent has quit IRC | 19:56 | |
| *** lukebrowning has quit IRC | 19:57 | |
| TheJulia | I can assure you, nobody wants to see my life. | 19:58 |
| *** lukebrowning has joined #openstack-tc | 19:59 | |
| harlowja | lol | 20:00 |
| harlowja | u could be famous (more famous!) | 20:00 |
| *** lukebrowning has quit IRC | 20:04 | |
| *** lukebrowning has joined #openstack-tc | 20:05 | |
| *** lukebrowning has quit IRC | 20:10 | |
| *** lukebrowning has joined #openstack-tc | 20:11 | |
| *** lukebrowning has quit IRC | 20:16 | |
| fungi | my lab is a camera-free zone for... er... security reasons. yeah, that's it | 20:17 |
| *** lukebrowning has joined #openstack-tc | 20:19 | |
| smcginnis | :) | 20:19 |
| mrhillsman | lol | 20:20 |
| fungi | harlowja: as for whether the foundation has a pr person/marketing manager responsible for trying to lead/figure out increasing visibility of the brand in non-usa regions, that's "lauren sell, vice president, marketing and community services" https://www.openstack.org/foundation/staff | 20:20 |
| fungi | and more specifically "heidi bretz, director of business development" | 20:21 |
| fungi | there are lots of additional people at the foundation working on various aspects of it too... claire, danny, allison, erin, anne, kendall, ashlee | 20:23 |
| *** lukebrowning has quit IRC | 20:23 | |
| fungi | (kendall w, not kendall n who does different awesome things unrelated to marketing) | 20:23 |
| fungi | roughly 40% of our full-time staff are marketing and business development, by my count | 20:25 |
| fungi | maybe more if you count the design and dev staff | 20:27 |
| fungi | er, design staff i meant | 20:27 |
| harlowja | cool | 20:41 |
| harlowja | the other root cause that might be useful to ummm, figure out | 20:42 |
| harlowja | is for lack of better terms, why devs. in various places hate/crap on openstack | 20:42 |
| harlowja | i've heard such things are happening more in various other communities ... | 20:43 |
| harlowja | ^ and i'd like to do some self-reflection and figure out how to stop/address it | 20:45 |
| *** lukebrowning has joined #openstack-tc | 20:48 | |
| *** lukebrowning has quit IRC | 20:53 | |
| *** lukebrowning has joined #openstack-tc | 21:00 | |
| *** diablo_rojo has quit IRC | 21:04 | |
| fungi | just like anywhere there's a perceived competition between technologies, you'll get loud "fans" of one making baseless disparaging remarks of whatever software they feel their favorite is competing with. for those detractors at least, there's not a lot to be done other than painting a smaller target on yourself | 21:05 |
| *** lukebrowning has quit IRC | 21:05 | |
| harlowja | what if we reached out to those fans and tried to digest there input? | 21:06 |
| fungi | cf, endless "pc vs mac" debates over the decadews | 21:06 |
| fungi | a lot of what i've digested is people repeating the same misinformation | 21:06 |
| *** lukebrowning has joined #openstack-tc | 21:06 | |
| harlowja | ok, to fix that we then need to do (?) | 21:07 |
| fungi | that's not to say that there aren't good criticisms out there, but they're generally drowned out by rabid fans who want to make themselves look like they're whoever's #1 fan by flinging derogatory insults at the rival team | 21:08 |
| fungi | so for example, we've all been putting a lot of effort into disambiguating openstack from other software projects in the eyes of the media and pundits in an attempt to remove ourselves from competition based on invalid comparisons | 21:09 |
| fungi | like the "you can use openstack and kubernetes together, here's how they're complimentary" sorts of articles and conference talks have been helping on that front, i think | 21:10 |
| *** lukebrowning has quit IRC | 21:12 | |
| fungi | basically we have to show that we're not in competition, we're in cooperation | 21:12 |
| *** lukebrowning has joined #openstack-tc | 21:13 | |
| fungi | getting more of the same people contributing to both openstack and those other projects helps it a lot too | 21:14 |
| harlowja | ya, i'm wondering if we have any close-person from the <other side> that we can bounce ideas off of | 21:14 |
| fungi | and cross-pollination between event attendees can also be good for that, as long as care is taken not to look like openstack people are invading some other community's event to "steal their users/devs" or whatever | 21:15 |
| harlowja | what about more cross-pollination between devs? | 21:15 |
| harlowja | just throwing this out there, switch all our channels to slack | 21:15 |
| fungi | you'll never get me doing open work through proprietary services | 21:16 |
| harlowja | i'm not saying i like it, i'm just trying to throw out things that (even if painful/not liked by everyone) could help | 21:16 |
| fungi | but using that suggestion for instance, we can certainly set an example of the virtuous cycle of using free software to develop free software | 21:16 |
| *** diablo_rojo has joined #openstack-tc | 21:17 | |
| *** lukebrowning has quit IRC | 21:17 | |
| harlowja | i get the view on principal, but at what point are we also willing to compromise for the health of ourselves and the 2 groups | 21:18 |
| *** lukebrowning has joined #openstack-tc | 21:22 | |
| dtroyer | It feels like there have been a number of suggestions lately for OpenStack (a previous hype-child) to change to be more like the current hype-child for some reason. It makes me feel like there is a lack of confidence in where we are or even WHO we are… Am I missing something? | 21:26 |
| harlowja | u are dean | 21:26 |
| harlowja | u are from planet earth | 21:26 |
| *** lukebrowning has quit IRC | 21:27 | |
| dtroyer | maybe the cyclical identity-crisis thing is just part of the nature of this level of decentralized development? I'm trying to find another project with similar cycles, and can't think of one that I have first-hand experience with | 21:28 |
| dtroyer | not that I have been in a lot of projects, but more than 2 or 3 | 21:28 |
| *** lukebrowning has joined #openstack-tc | 21:29 | |
| *** jroll has joined #openstack-tc | 21:33 | |
| flwang | hi there, is there any process to propose a community goal? | 21:33 |
| *** lukebrowning has quit IRC | 21:33 | |
| fungi | flwang: https://governance.openstack.org/tc/goals/#process-details | 21:34 |
| david-lyle | I think there's a bit of backlash from proprietary public clouds in the criticisms of OpenStack. Kubernetes was built to run on these proprietary providers originally (one specifically). It's a little open source in a much larger proprietary sandwich | 21:34 |
| *** kumarmn has quit IRC | 21:35 | |
| david-lyle | there's nothing wrong with that, it's just a different model | 21:35 |
| *** lukebrowning has joined #openstack-tc | 21:35 | |
| flwang | fungi: thanks | 21:36 |
| fungi | flwang: so basically, per that second paragraph, propose it as a goal in a thread on the openstack-dev ml, preferably at least including the [tc] subject tag | 21:36 |
| fungi | flwang: more formally, goal proposals end up as proposed patches to the goals directory in the openstack/governance repository | 21:37 |
| fungi | flwang: they're separated by cycle into directories here: http://git.openstack.org/cgit/openstack/governance/tree/goals | 21:38 |
| fungi | and you can copy and edit that template.txt to create a new one | 21:38 |
| fungi | you'll see the template has some instructions in it as placeholder text | 21:39 |
| fungi | david-lyle: agreed, the culture is going to be different in a community which is (or at least started out) inherently dependent on proprietary software/services vs one which was ground-up free software | 21:39 |
| *** lukebrowning has quit IRC | 21:40 | |
| fungi | i think that distinction is a big part of our community identity and something i wouldn't want to see us lose | 21:40 |
| fungi | it's something other communities struggle to achieve, which we tend to take for granted too often | 21:40 |
| *** lukebrowning has joined #openstack-tc | 21:41 | |
| mriedem | flwang: i bet i can guess what you're going to propose | 21:46 |
| flwang | mriedem: haha | 21:48 |
| mriedem | fungi: harlowja: btw, dims is working in both communities so he might have some insight | 21:48 |
| mriedem | also, -1 to slack; total distraction when we had to use it at ibm | 21:48 |
| flwang | good see you're around, buddy | 21:48 |
| flwang | mriedem: may I get your opinion about the idea before I put more effort on that? | 21:49 |
| mriedem | flwang: i don't even know what it is, | 21:49 |
| mriedem | but apparently i've say NO to it many many times before | 21:49 |
| mriedem | because i work in nova and that's all we do | 21:49 |
| mriedem | *said | 21:49 |
| flwang | mriedem: if so, i would stop putting effort if I can't get a support from Nova | 21:50 |
| flwang | mriedem: but meanwhile, i'm trying to figure out a way to collection notification from messaging bus | 21:51 |
| *** lukebrowning has quit IRC | 21:51 | |
| flwang | instead of talking to services hard to change | 21:51 |
| *** kumarmn has joined #openstack-tc | 21:52 | |
| *** lukebrowning has joined #openstack-tc | 21:52 | |
| mriedem | flwang: i'm being sarcastic, sorry | 21:53 |
| mriedem | why can't you just listen on the existing notifications that nova sends out? | 21:53 |
| dims | flwang : please do state your idea. matt is reacting to something on the mailing list :) | 21:53 |
| flwang | mriedem: it could be possible, but zaqar is not a place to listen the infra messaging bus | 21:54 |
| *** diablo_rojo has quit IRC | 21:54 | |
| flwang | we may have to leverage another service/process to collect all the notifications, like ceilometer | 21:55 |
| mriedem | flwang: i don't really know anything about zaqar | 21:55 |
| mriedem | are you suggesting something where nova has to push directly to zaqar REST APIs or something? | 21:55 |
| *** diablo_rojo has joined #openstack-tc | 21:55 | |
| flwang | mriedem: if you don't know anything about zaqar, why the idea is easily say no? | 21:55 |
| mriedem | flwang: i was being facetious | 21:56 |
| mriedem | because there have been several generic claims in that thread about how nova is generally unwilling to cooperate or work on anything that anyone ever wants, ever | 21:56 |
| mriedem | including things that, i at least, have never heard of | 21:57 |
| flwang | mriedem: if a lot people thing about that way, that could be true? | 21:57 |
| *** kumarmn has quit IRC | 21:57 | |
| *** kumarmn has joined #openstack-tc | 21:57 | |
| mriedem | if donald trump says everything on the news is fake, does that make that true too? | 21:57 |
| *** lukebrowning has quit IRC | 21:58 | |
| flwang | dims: thanks, I will talk with zaneb and any other people interested in that idea to see if we can propose that as a community goal | 21:58 |
| flwang | mriedem: we could be wrong, but that is out of the topic here | 21:58 |
| flwang | i'm just trying to see if we can figure out a way, let tenant user can get the resource notifications | 21:59 |
| *** lukebrowning has joined #openstack-tc | 21:59 | |
| flwang | for now, all notifications are sent to the infra message bus, there is no way for tenant user to consume them | 21:59 |
| flwang | that's the goal we want to address | 21:59 |
| flwang | i don't really care about the way, TBH | 21:59 |
| mriedem | ok. seems that zaqar (if that's the front end for the tenant user) could just listen on the notifications from the projects and queue those up for the tenant user to consume | 22:00 |
| flwang | mriedem: you're just proposing a change which could change the scope of zaqar | 22:00 |
| mriedem | but i don't know how you put policy controls in place for that, because a tenant user shouldn't know which host they are running on | 22:01 |
| mriedem | so we have a goal in mind, but no idea on how to implement it? | 22:01 |
| flwang | mriedem: depends on if we want to address it, we can define some policy for the sensitive notifications | 22:02 |
| mriedem | sorry i just thought this was something that's already been done in at least one project (heat) | 22:02 |
| flwang | to filter them out, i think | 22:02 |
| flwang | mriedem: yes, hence why we think if we can "copy" it to the other services | 22:02 |
| mtreinish | mriedem: or how do you isolate data to different tenants? You don't want to users in different tenants to have access to the notifications to each other's resources... | 22:03 |
| dims | flwang : is panko part of the possible solution? | 22:03 |
| *** lukebrowning has quit IRC | 22:04 | |
| mriedem | mtreinish: right, that's why i assumed something would be storing the notifications i guess, and those notifications would need some user/project context information | 22:04 |
| flwang | dims: could be, but I' | 22:04 |
| zaneb | right, so the idea is not to let users read the rabbit bus | 22:04 |
| flwang | i'm not sure if Heat is doing any special things | 22:05 |
| zaneb | the idea is to allow users to request notifications about their guest | 22:05 |
| flwang | zaneb: ^ | 22:05 |
| *** lukebrowning has joined #openstack-tc | 22:05 | |
| mriedem | zaneb: where do they make the request for this if heat is doing it? the heat api? | 22:05 |
| flwang | zaneb: is there any difference b/w the usage in heat for zaqar and put something in panko to address this | 22:06 |
| flwang | ? | 22:06 |
| mriedem | or zaqar? | 22:06 |
| zaneb | one way, and what we have done in Heat, is to allow the user to specify a list of Zaqar queues when they create a stack, and then when stuff the user is interested in happens, Heat posts notifications to those queues | 22:06 |
| zaneb | something similar in Nova would be cool, but would involve changing the Nova API and we've been told (most recently by johnthetubaguy) that that is off-limits for all time | 22:07 |
| mriedem | so nova would need to post notifications to the zaqar rest api? | 22:07 |
| mriedem | how hard is it to just swap out the notification driver in oslo.messaging to be a zaqar driver that pushes to the zaqar rest api? | 22:07 |
| mriedem | rather than the rpc queue | 22:07 |
| zaneb | the alternative obviously is to write a separate service that nobody will ever install to read messages from the (unreliable) rabbit queue, filter them, and proxy them to the requested Zaqar queues | 22:08 |
| dims | zaneb : flwang : yes, you can write a notification driver for zaqar and specify that in nova.conf (https://github.com/openstack/oslo.messaging/tree/master/oslo_messaging/notify) | 22:08 |
| flwang | dims: zaqar is not at that level unfortunately | 22:08 |
| flwang | flaper87: ^ | 22:08 |
| mriedem | so does heat talk to zaqar over rest or rpc? | 22:09 |
| dims | only question is ... is the set of events from nova, is that enough | 22:09 |
| zaneb | dims, mriedem: that part is not hard; the hard part is the interface to allow the user to configure it (i.e. what queue(s) it goes to) | 22:09 |
| zaneb | mriedem: ReST | 22:09 |
| dims | zaneb : that would not be upto nova right? something else has to do that | 22:09 |
| mriedem | dims: i assume you'd have to create the vm with a specific queue(s) | 22:09 |
| *** lukebrowning has quit IRC | 22:10 | |
| zaneb | dims: what mriedem said | 22:10 |
| mriedem | and nova would be responsible for pushing notifications to the zaqar rest api for said queues | 22:10 |
| zaneb | exactly | 22:10 |
| dims | mriedem : you can do that in the code that you write for notification driver | 22:10 |
| dims | no need to do that in Nova | 22:10 |
| mriedem | dims: you'd still have to pass the queue from the instance to the driver | 22:11 |
| mriedem | like, nova sends a big ass notification payload for an instance today, | 22:11 |
| zaneb | dims: the user doesn't interact with the notification driver, they interact with the Nova API. So the Nova API has to accept the list of queues | 22:11 |
| mriedem | which includes things like block devices and ports | 22:11 |
| mriedem | that would also have to contain zaqar queues | 22:11 |
| dims | we can make sure that the data the driver gets has enough info to figure out which instance ... | 22:11 |
| *** lukebrowning has joined #openstack-tc | 22:11 | |
| dims | this is class separation of duty ... nova's duty is to put out enough info in the hooks | 22:12 |
| mriedem | so POST /servers/ {'queues': [1,2,3]} -> pass that through to the notification driver, which reads the queues off the payload and posts to the zaqar API | 22:12 |
| mriedem | dims: what hooks? | 22:13 |
| dims | notification driver entry point + events from various spots in nova | 22:13 |
| mriedem | but they are saying the user has to create the server with specific queues | 22:13 |
| dims | mriedem : why would we want to hard code zaqar queues in nova? | 22:13 |
| mriedem | i'm not talking about hard-coded queues | 22:13 |
| mriedem | i assume this is like creating a server with a specific volume and ort | 22:14 |
| mriedem | *port | 22:14 |
| flwang | dims: there is no hardcode | 22:14 |
| flwang | mriedem: correct | 22:14 |
| mriedem | you create a server with a specific volume, port, image and queue | 22:14 |
| dims | listening | 22:14 |
| mriedem | and the instance payload that goes to the notification driver contains that queue entry, | 22:14 |
| mriedem | so the zaqar driver knows which queue to post the notification message to for that instance | 22:15 |
| zaneb | (in a perfect world you can probably add and remove queues later as well) | 22:15 |
| mriedem | https://docs.openstack.org/nova/latest/reference/notifications.html#versioned-notification-samples | 22:15 |
| dims | like an annotation .. if it is present, pass that through to the notification driver | 22:15 |
| zaneb | yeah | 22:16 |
| mriedem | what about just passing this through as specific metadata? | 22:16 |
| *** lukebrowning has quit IRC | 22:16 | |
| mriedem | the instance metadata is passed through the notification payload | 22:16 |
| mriedem | and you can change metedata on the fly | 22:16 |
| dims | ++ mriedem | 22:16 |
| mriedem | b/c of the metadata service | 22:16 |
| mriedem | or tags | 22:16 |
| mriedem | instance tags | 22:16 |
| mriedem | same idea | 22:16 |
| mriedem | create a server with queue:foo as metadata | 22:17 |
| flwang | mriedem: it works i think | 22:17 |
| mriedem | zaqar notification driver reads from that | 22:17 |
| zaneb | I could see that working | 22:17 |
| mriedem | cinder has a similar concept | 22:17 |
| mriedem | neutron has tags | 22:17 |
| mriedem | not sure about meta for neutron | 22:17 |
| mriedem | but you could do a queue=foo tag | 22:17 |
| dims | so that's what i was saying we have the pieces there already | 22:17 |
| flwang | mriedem: but I don't think there could be a oslo.messaging driver | 22:17 |
| mriedem | can i go back to saying no to everything else now? :) | 22:17 |
| mriedem | flwang: why? | 22:18 |
| flwang | zaneb: is it possible to exact the code in heat to make it like a common lib? | 22:18 |
| *** lukebrowning has joined #openstack-tc | 22:18 | |
| dims | flwang : please don't ask Nova to depend explicitly on zaqar | 22:18 |
| flwang | mriedem: zaqar is a messaging service, not a message bus/lib like rabbitmq/qpid | 22:18 |
| zaneb | dims: it'd be a soft dependency, obviously, if there was one | 22:18 |
| dims | we already have the versioned notifications through notification driver. that's the place to plug in | 22:19 |
| mriedem | flwang: it doesn't have to be | 22:19 |
| mriedem | flwang: i can write a notification driver that dumps everything to a text file | 22:19 |
| flwang | dims: no hard dependency for sure | 22:19 |
| dims | and another process picks up from the file and writes to zaqar | 22:19 |
| dims | which is essentially what aodh does | 22:19 |
| mriedem | or zaqar just has a service that listens on the rpc notifications, like searchlight, ceilometer, designate, that other 'm' project that does task orchestration | 22:20 |
| dims | mriedem : right-o | 22:21 |
| mriedem | mistral? | 22:21 |
| mriedem | yup | 22:21 |
| flwang | mriedem: personally, i think that's the problem openstack need to address | 22:21 |
| mriedem | yeah so this all seems totally possible using existing APIs for nova, cinder and neutron, not sure about others | 22:21 |
| dims | flwang : the pieces are there, someone needs to stitch it together | 22:21 |
| flwang | i don't want to see "all' projects write the dup code to listen the MB | 22:22 |
| mriedem | as far as the user being able to create a resource with a specific key that the listening service is looking for | 22:22 |
| dims | flwang : so do that once in zaqar ... or use what is in Aodh | 22:22 |
| flwang | and I know the story of mistral, why they have to do that, which they don't want to | 22:22 |
| mriedem | flwang: what are you proposing? that nova has specific points where it calls every other project's REST API for whatever information they need to get from nova? | 22:23 |
| *** lukebrowning has quit IRC | 22:23 | |
| mriedem | creating a common library is one thing | 22:23 |
| zaneb | flwang: I'm not sure there's enough code to bother with a common lib: http://git.openstack.org/cgit/openstack/heat/tree/heat/engine/clients/os/zaqar.py#n87 | 22:23 |
| mriedem | if that's the major problem, then why don't the people that feel strongly about that problem work on the library? | 22:23 |
| flwang | i will seriously consider if we can create a tiny process/service in zaqar to do that, like therve has done https://github.com/therve/osn/tree/master/osn | 22:23 |
| *** david-lyle has quit IRC | 22:24 | |
| *** lukebrowning has joined #openstack-tc | 22:24 | |
| mriedem | this is the designate code for consuming notifications i think https://github.com/openstack/designate/blob/master/designate/sink/service.py | 22:24 |
| flwang | mriedem: probably that's the last easy sort and no body can block them | 22:24 |
| flwang | mistral was blocked almost 1 year and they failed, Mistral PTL, pls correct me if I'm wrong | 22:25 |
| dims | flwang : https://github.com/openstack/oslo.messaging/blob/master/tools/simulator.py#L262 | 22:26 |
| mriedem | blocked on what? | 22:26 |
| dims | flwang : that's the code we use to test both rpc and notifications and simulate load etc | 22:27 |
| flwang | mriedem: can't remember the details, but you know the result now, they have to listen MB themselves | 22:27 |
| dims | flwang : the design is either write a driver or listen to events .. if the events are not rich enough then propose more events that we can add. what can we change here? | 22:28 |
| mriedem | flwang: ok i don't know what they expected nova to do | 22:28 |
| flwang | mriedem: no,it's not related to nova, sorry | 22:29 |
| mriedem | right, i mean, seriously, there are existing APIs that have been around for years that can be used to do this thing that apparently nova is blocking everyone on | 22:29 |
| *** lukebrowning has quit IRC | 22:29 | |
| dims | mriedem : don't go there :) | 22:29 |
| zaneb | one other aspect of this that comes up a lot is that we'd need to sanitise the notifications to strip out stuff that is not for end-users' eyes. there are 3 or 4 different projects tackling this or wanting to tackle it independently (Aodh, Mistral, Searchlight off the top of my head) | 22:29 |
| zaneb | I wonder if there is a way we can mark fields in the notification as sensitive or not at the source | 22:30 |
| mriedem | dims: have you read the 1-year dev cycle thread? | 22:30 |
| mriedem | like, all of it? | 22:30 |
| zaneb | if we could do that then I could see this being implemented as a oslo-messaging driver | 22:30 |
| *** lukebrowning has joined #openstack-tc | 22:30 | |
| mriedem | zaneb: what is sensitive depends on the user context and policy | 22:30 |
| dims | mriedem : yep, i hear your frustration. | 22:31 |
| flwang | zaneb: thanks, that's one of the reasons, zaqar can't perfectly handle this case | 22:31 |
| zaneb | mriedem: maybe mark the things that are definitely-not-sensitive then? ;) | 22:31 |
| flwang | mriedem: like the host info | 22:31 |
| mriedem | rather than completely rely on just passing the nova notification payload through to the end user, | 22:31 |
| mriedem | why wouldn't the consuming service just have a mapping of what it exposes to the end user | 22:32 |
| mriedem | and that is based on context and policy in that service | 22:32 |
| mriedem | so if zaqar wants to expose instance.delete.end, fine, but it could also pull the things out of the payload that it cares about and discard the rest | 22:32 |
| zaneb | zaqar is a dumb transport | 22:33 |
| flwang | mriedem: who maintain the mapping? | 22:33 |
| flwang | zaqar don't know anything about the notifications | 22:33 |
| mriedem | flwang: the consuming project that is ultimately exposing things out of it's REST API to the tenant user | 22:33 |
| flwang | the content i mean | 22:33 |
| mriedem | searchlight is already doing this | 22:33 |
| smcginnis | Definitely doesn't seem like Nova's responsibility to know either. | 22:33 |
| zaneb | zaqar is just a way of getting messages to the user and letting them authenticate with keystone | 22:33 |
| mriedem | so cinder already has a messages API | 22:34 |
| mriedem | seems we don't even need zaqar then if we're all just going to implement a messages API in each service | 22:34 |
| zaneb | smcginnis: I disagree. Nova knows what data is e.g. already exposed in the API and what is internal, and it is well placed to be a single source of truth about that | 22:35 |
| *** lukebrowning has quit IRC | 22:35 | |
| flwang | it's surprised that Nova don't know what notifications are sensitive | 22:35 |
| mriedem | flwang: i'm sure we could audit all of that, but again, it's a policy and per-context thing | 22:36 |
| mriedem | and you'd be talking about duplicating that across all projects that you need this capability from | 22:36 |
| *** lukebrowning has joined #openstack-tc | 22:36 | |
| zaneb | mriedem: let me put it another way... if Nova doesn't know then there is no way the other 3 or 4 or more services that want to expose this stuff will ever figure it out correctly | 22:37 |
| smcginnis | Well, like mriedem said, seems like it depends on context and policy what is sensitive and what is not. | 22:38 |
| flwang | mriedem: do we have a policy can control the notification's sensitivity ? | 22:38 |
| mriedem | zaneb: that's why i was suggesting a mapping so the consumer can pick what they choose to expose and to whom | 22:38 |
| zaneb | the current situation is outright dangerous - see my comments on https://review.openstack.org/#/c/308664/ as an example | 22:38 |
| mriedem | flwang: no, because notifications are all assumed to be over rpc and internal | 22:38 |
| flwang | mriedem: true | 22:39 |
| zaneb | mriedem: consumer, or operator? | 22:39 |
| flwang | zaneb: thanks for the link which is the one i mentioned above | 22:40 |
| flwang | either consumer or even operator, i don't think they have the knowledge to maintain the mapping | 22:41 |
| mriedem | zaneb: nova has no idea who or what is consuming it's notifications | 22:41 |
| mriedem | and how those payloads are going to be used, whether they are exposed to a tenant user or not | 22:41 |
| *** lukebrowning has quit IRC | 22:42 | |
| smcginnis | Or some internal service that actually needs that "sensitive" information. | 22:42 |
| mriedem | right, | 22:42 |
| flwang | mriedem: yep, out of curious, do we have a cross project payload format ? so that we can define some sensitive format for fieleds like we're doing for logs | 22:42 |
| mriedem | like, something listens on service.disabled | 22:42 |
| mriedem | service.update or whatever it is | 22:42 |
| mriedem | flwang: no there is no standard payload for notifications in openstack | 22:43 |
| *** lukebrowning has joined #openstack-tc | 22:43 | |
| flwang | mriedem: so it's mess world | 22:43 |
| zaneb | well, I'd phrase that as: it is assumed that the only things listening are those specifically allowed by the operator, and they have operator-level privileges | 22:43 |
| mriedem | yes i would agree with that | 22:43 |
| dims | back in a bit, will read backlog :) | 22:44 |
| mriedem | also, going back to a priority question, how many users are asking for this given i've never heard of it before today? | 22:45 |
| mriedem | that is probably a harsh question, | 22:45 |
| zaneb | so if we want to give that data to the user, either the service needs to decide which bits are safe to export to user-space, or the operator does (via some configuration), or some combination of the two | 22:45 |
| mriedem | but we have to prioritize things based on limited resources | 22:45 |
| flwang | mriedem: that probably because nova is at the lower level than heat, zaqar.... | 22:46 |
| zaneb | mriedem: Heat and Horizon have been wanting to get this stuff from notifications instead of polling since... the beginning of time | 22:46 |
| mriedem | what is "this stuff" specifically? | 22:46 |
| mriedem | rbac controlled notification payload content? | 22:47 |
| zaneb | when events happen (e.g. server build completes) | 22:47 |
| flwang | get the notifications I'm interested in | 22:47 |
| mriedem | zaneb: and notifications have been around since the beginning of time | 22:47 |
| mriedem | so what am i missing besides the rbac thing and hooking in zaqar somehow? | 22:48 |
| *** lukebrowning has quit IRC | 22:48 | |
| mriedem | which the latter could all be done already | 22:48 |
| zaneb | actually we've wanted this since before notifications were a thing. but we can't have everyone drinking from that firehose | 22:48 |
| zaneb | the other thing though is that it's important for autonomous applications to export the events to userspace | 22:49 |
| *** lukebrowning has joined #openstack-tc | 22:49 | |
| zaneb | flwang and I worked on a thing where you can trigger Mistral workflows from a zaqar message | 22:50 |
| zaneb | so applications can manage themselves (e.g. recover from a failure) in user-controlled ways, if they have access to events | 22:50 |
| flwang | we're trying to provide more for tenant user besides opertors | 22:51 |
| mriedem | yup i get that | 22:52 |
| zaneb | and a large fraction of users are ignoring OpenStack and going straight to AWS/Google/Azure because they have this and we don't | 22:52 |
| flwang | that's the another topic maybe out of this, which i struggle is, if openstack should do more innovation above the infra layer | 22:53 |
| flwang | zaneb: and I think yes, but somebody may have different opinion | 22:53 |
| *** lukebrowning has quit IRC | 22:54 | |
| mriedem | so there are 2 extreme sides to this right? on one end, zaqar today can build something to listen on notifications and map those to things it wants to expose to it's users, using some rbac on the zaqar side, | 22:54 |
| mriedem | and on the extreme, nova/cinder/glance/neutron/keystone/everyone else standardizes on a way to have rbac / context based notification payloads so that zaqar can just pass those through to the end user w/o having to worry about exposing sensitive information | 22:54 |
| mriedem | do i understand those two extremes correctly? | 22:55 |
| *** lukebrowning has joined #openstack-tc | 22:55 | |
| zaneb | mriedem: in the first one it's unlikely to be Zaqar itself, more likely Aodh or some yet-to-be-created service. other than that, I think that's broadly accurate | 22:56 |
| *** david-lyle has joined #openstack-tc | 22:56 | |
| mriedem | ok, so having said that, | 22:56 |
| mriedem | which do you think would get one faster? | 22:56 |
| mriedem | *done | 22:57 |
| mriedem | also considering that nova is, i think, the only project that has versioned notifications right now | 22:57 |
| mriedem | versioned payloads might not matter as long as projects never ever remove keys from the payload | 22:58 |
| zaneb | tbh I don't much care about the events from any of those other projects | 22:58 |
| mriedem | you don't, but do your users? | 22:58 |
| mriedem | or the users of this thing | 22:59 |
| mriedem | i have no idea what users of this thing want, they don't talk to me | 22:59 |
| zaneb | horizon/heat would probably care about some of them | 22:59 |
| zaneb | userspace users I think would mostly care about servers | 22:59 |
| flwang | afk for a while, will be back later, thank you mriedem and zaneb | 22:59 |
| mriedem | alright, well at least i understand the use case a bit more | 23:00 |
| *** lukebrowning has quit IRC | 23:00 | |
| zaneb | to answer you question though, it's probably similar | 23:00 |
| mriedem | which is 100% more than i did when i read in the ML that "Sure, it would be dead easy, but Nova cores have made it abundantly clear that under no circumstances would they ever accept any code like this in Nova. Hence it's on this list. " | 23:01 |
| zaneb | the first is a filter-the-firehose problem | 23:01 |
| zaneb | the second is mostly a configuration problem | 23:01 |
| *** lukebrowning has joined #openstack-tc | 23:02 | |
| zaneb | yes, sorry, that was a fairly glib comment even by my standards | 23:02 |
| *** diablo_rojo has quit IRC | 23:02 | |
| zaneb | mriedem: etherpad from the last discussion I recall, FWIW: https://etherpad.openstack.org/p/newton-alternatives-to-polling | 23:03 |
| mriedem | "I don't think Nova should directly integrate to zaqar. It seems like it should be driver. oslo.notification. Lots of projects will need these kinds of hooks." | 23:04 |
| mriedem | ^ exactly what dims and i were saying | 23:04 |
| mriedem | "Either Nova emits new notifications, or a service massages current notifications to make them user consumable." | 23:05 |
| mriedem | exactly what i was saying | 23:05 |
| mriedem | so solutions have been around for a couple of years now | 23:05 |
| mriedem | but no one worked on it because john said nova was'nt going to push directly to zaqar/ | 23:06 |
| mriedem | ? | 23:06 |
| *** lukebrowning has quit IRC | 23:07 | |
| zaneb | it's more complicated than that | 23:07 |
| zaneb | so the stuff in the actions (basically Heat prototyping this) did happen | 23:08 |
| *** lukebrowning has joined #openstack-tc | 23:08 | |
| zaneb | the 4th bullet that you reference there is problematic, it creates a huge load on whatever has to process/filter the firehose | 23:09 |
| mriedem | zaneb: someone has to do that somewhere yes? | 23:09 |
| mriedem | be it nova or the consuming service | 23:09 |
| zaneb | if it's filtered at the source then it's not so bad | 23:10 |
| mriedem | well you have to define the filters then | 23:10 |
| mriedem | which also seems complicated | 23:10 |
| zaneb | therve did build a prototype of the second bullet point, which I think flwang referred to earlier | 23:10 |
| zaneb | mriedem: well, you say that for this server that I'm booting here are the queues I want to post to. That saves something else from having to listen to every notification for every server in every tenant and figure out which if any queues it needs to proxy them to | 23:12 |
| mriedem | zaneb: but i pointed out you can do that with server metadata using a specific key (or a server tag) and make the (yet non-existent) oslo.messaging notification driver handle that | 23:13 |
| mriedem | the metadata gets passed all the way through | 23:13 |
| *** lukebrowning has quit IRC | 23:13 | |
| mriedem | and is mutable | 23:13 |
| zaneb | yeah, that's a good idea and one that did not come up in the session to the best of my recollection | 23:13 |
| zaneb | the idea that came up was the first bullet point, and johnthetubaguy made it clear that any api change was off the table | 23:14 |
| *** lukebrowning has joined #openstack-tc | 23:14 | |
| *** diablo_rojo has joined #openstack-tc | 23:14 | |
| zaneb | to be honest, a big part of the dynamic is that it's very difficult to get people to install a new service. and if it's not there on a critical mass of installations, then you can't write an autonomous application that's interoperable between OpenStack clouds so people won't bother | 23:16 |
| zaneb | so nobody was super motivated to work on the second bullet point because it looked like a dead end in terms of adoption | 23:17 |
| mriedem | sure, and that sucks, | 23:17 |
| mriedem | but it also sucks to put all knobs into nova | 23:17 |
| mriedem | because we suck managing all those knobs | 23:17 |
| mriedem | i just wish people would also consider that rather than think nova is only ever saying no or "not right now" to things, rather than we're just saying no because we like it | 23:18 |
| zaneb | it's my contention that if we had had this in Nova from very early on, we could have avoided adding a lot of even-harder-to-manage knobs like live migration, because we could have positioned OpenStack as a platform for cloud-aware applications instead of a VMWare replacement | 23:19 |
| *** lukebrowning has quit IRC | 23:19 | |
| zaneb | but that's a whole other philosophical discussion :) | 23:19 |
| mriedem | idk, isn't part of the "nova/cinder/glance are way too tightly coupled and all have to be installed together" mentality in part because they started in a single project? | 23:20 |
| mriedem | anyway, i've gotta run | 23:21 |
| mriedem | i would just like people to know that we're not working on all the things not because we like to say no or because we're bums not doing anything, but we also have limited resources and have to prioritize what it is we work on | 23:22 |
| *** lukebrowning has joined #openstack-tc | 23:24 | |
| *** lukebrowning has quit IRC | 23:29 | |
| *** lukebrowning has joined #openstack-tc | 23:30 | |
| *** lukebrowning_ has joined #openstack-tc | 23:32 | |
| *** lukebrowning has quit IRC | 23:34 | |
| *** lukebrowning_ has quit IRC | 23:37 | |
| *** lukebrowning has joined #openstack-tc | 23:38 | |
| *** kumarmn has quit IRC | 23:46 | |
| *** lukebrowning has quit IRC | 23:48 | |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!
