Monday, 2025-09-08

*** mhen_ is now known as mhen02:00
*** trident- is now known as trident09:49
tkajinamo/ I wonder if update of metadata can be accepted as an exception during FF period ? The nova feature was merged at the last-minute before its FF data and I'm hoping to get this merged to glance for consistency. https://review.opendev.org/c/openstack/glance/+/95998410:25
tkajinamupdate of *metadef* I mean10:25
dansmithI defer to the PTO but metadefs are not code so it seems (IMHO) fine to lump them in with bug fixes and test changes, up to rc113:49
dansmithPTO? PTL..maybe I need some PTO :D13:58
frickleriiuc the PTL is on PTO, that might explain the confusion ;)14:02
dansmithabhi is, but I didn't think croelandt was (the former is the PTL)14:04
croelandttkajinam: we're not doing rc1 iirc15:26
croelandtoh no actually we are15:29
croelandtyeah OK i can probably merge this and release rc1 then15:29
tkajinamyeah rc1 is supposed to be released in a few days15:31
croelandtdansmith: does the patch look good to our resident Nova expert?15:32
croelandttkajinam: not sure why the reference to KVM disappeared15:32
croelandtdoes AMD SEV replace KVM in Nova or smoething?15:32
tkajinamno. AMD SEV is used via qemu+KVM15:33
dansmithcroelandt: tkajinam is your resident nova expert for this feature :D15:34
tkajinamX-D15:34
tkajinamcroelandt, https://review.opendev.org/c/openstack/nova/+/78143715:34
tkajinamupdate of the link follows that doc restructure merged a few years back15:34
croelandttkajinam: but can you use KVM without AMD SEV?15:35
tkajinamyes15:35
croelandtI wonder if it makes sense to have KVM mentioned for hw_mem_encryption then15:35
tkajinamAMD SEV is a kind of additional feature15:35
croelandtsee what I mean?15:35
tkajinamah, ok15:35
croelandtlike, let's say I care about KVM but don't know about AMD SEV15:36
croelandtI'm going to be a bit confused here15:36
croelandtso maybe we mention the docs for AMD SEV in the hw_mem_encryption_model section15:36
croelandtbut we keep a link to KVM-specific doc for hw_mem_encryption?15:36
croelandtdoes that make sense?15:37
opendevreviewTakashi Kajinami proposed openstack/glance master: Add new metadata definition of memory encryption model  https://review.opendev.org/c/openstack/glance/+/95998415:38
tkajinamcroelandt, does this match your expectation ? https://review.opendev.org/c/openstack/glance/+/959984/3..4/doc/source/admin/useful-image-properties.rst15:38
croelandtnice, when I lazily grep for KVM I'll find what I'm looking for ;)15:40
croelandt\o/15:40
tkajinamcroelandt, nova no longer maintains SEV related descriptions in KVM section so we can no longer use that kvm specific link15:40
tkajinamso I mentioned kvm in the description, instead15:40
croelandtok, I see15:40
croelandttkajinam: but what about https://docs.openstack.org/nova/latest/admin/configuration/hypervisor-kvm.html ? Couldn't  this be part of the section about hw_mem_encryption?15:42
croelandtoh no ok hw_mem_encryption is not even in KVM15:42
tkajinamyeah it was migrated to the top-level dedicated doc for SEV15:43
croelandtyeah but also technically, hw_mem_encrpytion only makes sense if you use SEV15:43
tkajinamyup15:45
tkajinam... until we add support for Intel TDX :-P15:45
tkajinam(TDX is similar to SEV, but is available in Intel CPUs. Ignore it now.15:46
croelandtApparently consumers have been ignoring Intel CPUs for a little while :D15:47
croelandtok so I've pushed it, and I'll release rc1 once it's in15:49
opendevreviewMarkus Hentsch proposed openstack/glance master: Standardization of encrypted images  https://review.opendev.org/c/openstack/glance/+/92629515:51
tkajinamcroelandt, thanks !16:28
opendevreviewMerged openstack/glance master: Add new metadata definition of memory encryption model  https://review.opendev.org/c/openstack/glance/+/95998420:50

Generated by irclog2html.py 4.0.0 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!