| opendevreview | Taketani Ryo proposed openstack/nova master: mem-enc: stop using _get_mem_encryption_config() for SEV checks https://review.opendev.org/c/openstack/nova/+/967970 | 07:28 |
|---|---|---|
| opendevreview | Taketani Ryo proposed openstack/nova master: mem-enc: refactor memory encryption trait logic for extensibility https://review.opendev.org/c/openstack/nova/+/967971 | 07:28 |
| opendevreview | Taketani Ryo proposed openstack/nova master: mem-enc: make RP creation independent of specific encryption models https://review.opendev.org/c/openstack/nova/+/967972 | 07:28 |
| opendevreview | Taketani Ryo proposed openstack/nova master: mem-enc: refactor _guest_configure_mem_encryption() for extensibility https://review.opendev.org/c/openstack/nova/+/967973 | 07:28 |
| opendevreview | Taketani Ryo proposed openstack/nova master: mem-enc: adjust requirement checks for mem_encryption guests https://review.opendev.org/c/openstack/nova/+/967974 | 07:28 |
| opendevreview | Taketani Ryo proposed openstack/nova master: mem-enc: introduce a check between mem_encryption and locked_memory https://review.opendev.org/c/openstack/nova/+/971300 | 07:28 |
| opendevreview | Taketani Ryo proposed openstack/nova master: mem-enc: address minor issues pointed out in the review https://review.opendev.org/c/openstack/nova/+/983278 | 07:28 |
| TheJulia | sean-k-mooney[m]: quick comment looking at the back scroll, there is an interesting and confusing shift we've seen when the topic of billing baremetal has come up, in that I think most take entirely different views on approaches at that cross-over to whole hosts, likely rooted in the shift from op-ex accounting of expenditures and already sunk cost against the prior cap-ex versus the sunk cap-ex cost of just the gear and | 13:24 |
| TheJulia | then the direct tie to op-ex of use. Unfortunately, the venn diagram of the topic has never really yielded solid consensus from the baremetal heavy operators that I've spoken to. :( | 13:24 |
| sean-k-mooney | TheJulia: :) ist more that we recommend encoding model sepcific informationin resouce classes rathre then traits if the reosuce shoudl not be consider interchangable. i.e. if you have 2 dell server skus and form a billing/perfoamce/consumtion point of view you treat them the same because objectivly an end user would not care whic of the two host they got then its ok to share | 13:28 |
| sean-k-mooney | a resouce class but if they are not interchangable they should use diffent resouce classes | 13:28 |
| sean-k-mooney | so the Standard baremental one shoudl really only be used if you only have 1 type of server. | 13:28 |
| sean-k-mooney | or one tyep in a given AZ or regions for example | 13:28 |
| TheJulia | oh, yeah, absolutely, I guess what I'm trying to get across, is people seem to really stop caring about billing baremetal at some point and its weeeeeird. | 13:29 |
| sean-k-mooney | oh ya | 13:29 |
| sean-k-mooney | well at some pont it just power cost as well right | 13:29 |
| TheJulia | OR, they have entirely different models built into their budgets and just roll with that | 13:29 |
| sean-k-mooney | i.e. you run them till they die or you run out of power/rack space | 13:29 |
| sean-k-mooney | ya that too | 13:29 |
| TheJulia | well, power + licenses + support on some level, but yeah | 13:29 |
| TheJulia | oh, and cooling | 13:30 |
| TheJulia | when I ran data centers, getting leadership to accept budgetting for cooling repairs or even improvements was quite literally the hardest problem even with temp sensors, load graphs, power curve and usage calculations | 13:30 |
| TheJulia | I never quite understood that resistance to keeping the server nice and happy at ideal temps | 13:31 |
| sean-k-mooney | i was an intern at micosoft on the office infra engeieing team and i remmeber the day the sent out a email saying you have 10 minths ot shutdown allw workloads in a data center | 13:31 |
| sean-k-mooney | they loss all redunant cooling in the building and the rack room hat already pass 48c | 13:31 |
| TheJulia | yup, nothing quite like walking into a data center at 61-62c. | 13:32 |
| sean-k-mooney | there was some auto cutoff on the power that gong to dump the racks off line if it got to like 55 or something that they were trying to get ahead of | 13:32 |
| TheJulia | (we lost all cooling on christmas day once, it was awesome!) | 13:33 |
| sean-k-mooney | i bet | 13:33 |
| sean-k-mooney | fyi ironic with devstack does not install on debina 13 due to a missing depency that was replaced | 13:33 |
| TheJulia | yeah, the transformers trip was at 65c | 13:33 |
| sean-k-mooney | shim-signed i think | 13:34 |
| TheJulia | sean-k-mooney: has it been logged in launchpad? | 13:34 |
| TheJulia | oh, figures | 13:34 |
| sean-k-mooney | https://paste.opendev.org/show/bnsVQDVkNQ16t6PvmTVr/ | 13:35 |
| sean-k-mooney | i can report that quickly. i kicked off an install yesterday and i was just checking on it im going to kick it off on ubuntu again | 13:36 |
| sean-k-mooney | TheJulia: i assuem the ironic launchpad this is related i think ti the pxe image install or somethign related to that | 13:36 |
| sean-k-mooney | https://github.com/openstack/ironic/blob/master/devstack/lib/ironic#L726 | 13:38 |
| TheJulia | yeah, because of all the things around shim packaging we've seen those names change/evolve a lot over the years, that should really be in bindep :\ | 13:39 |
| TheJulia | ... and LP doesn't load for me today. Joy! | 13:39 |
| sean-k-mooney | hum same | 13:40 |
| sean-k-mooney | it might just be very slow it did fineally load partly with a timeout error | 13:40 |
| TheJulia | took close to 90 seconds on my end, but yeah, did finally load | 13:41 |
| TheJulia | 30+ seconds to just sort the bug list, awesome! | 13:42 |
| sean-k-mooney | shall i file it or are you on it | 13:42 |
| * TheJulia clicks "report a bug" | 13:42 | |
| TheJulia | I'll do it | 13:42 |
| sean-k-mooney | cool | 13:42 |
| sean-k-mooney | im seeing up a dev env badsed on the patch jay created so if i get time to look at the startup performance between now and the ptg i have a place to test things out | 13:43 |
| sean-k-mooney | then you on im goign to enjoy the day off :) | 13:43 |
| TheJulia | enjoy! | 13:50 |
| TheJulia | gibi: I don't see any issue with your tls socket change, that being said it looks like somewhere the job setup for graphical console in ironic ci fell down or got lost, I'm going to sync up with stevebaker next week and we'll get it on the top of our list to sort out next week. | 14:17 |
| Uggla | FYI I have booked the PTG slots here: https://ptg.opendev.org/ptg.html. This will happen mostly in Liberty room. Note this is currently draft, I will arrange when I'll be back from PTO. | 15:55 |
| opendevreview | melanie witt proposed openstack/nova master: TPM: fixup for bump service version documentation https://review.opendev.org/c/openstack/nova/+/978836 | 22:56 |
| opendevreview | melanie witt proposed openstack/nova master: TPM: support live migration of `user` secret security https://review.opendev.org/c/openstack/nova/+/983333 | 22:56 |
| opendevreview | Merged openstack/nova master: Add nova-vtpm job https://review.opendev.org/c/openstack/nova/+/978841 | 23:51 |
Generated by irclog2html.py 4.1.0 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!