opendevreview | Clark Boylan proposed openstack/diskimage-builder master: Setup C.UTF-8 as a RPM install lang in yum-minimal https://review.opendev.org/c/openstack/diskimage-builder/+/930932 | 16:07 |
---|---|---|
clarkb | that change passes testing but I'm not in a good spot to do a local build to confirm it works myself. It seems low risk if the build succeeds (because that implies we can still install packages successfully) and worst case we just won't have the locale installed as expected which is the status quo | 17:09 |
clarkb | but maybe some of yall more familiar with centos and rpm can take a look and decide if that is safe as is | 17:09 |
opendevreview | Julia Kreger proposed openstack/diskimage-builder master: Reduce LVM extent usage https://review.opendev.org/c/openstack/diskimage-builder/+/930950 | 20:56 |
TheJulia | so ^^^ is a stupidly fun one and I'd really <3 some reviews of it | 20:57 |
TheJulia | bottom line, we tried to run growvols on a 4k disk and stuff exploded | 20:57 |
TheJulia | and I think it is all boiling down to extent alignmnet | 20:57 |
JayF | we should align the extents of the hardware engineers who decided to buck decades of block-size-standard ;) | 20:58 |
JayF | okay, I'm a little outta my depth reviewing that, I get the idea and I think you're on the right track. It makes sense | 20:59 |
TheJulia | I support this, in the lightest most gentle way | 21:09 |
TheJulia | I guess I'm fighting this question if I should leave at least one extra extent | 21:10 |
TheJulia | since it can line up perfectly now | 21:10 |
TheJulia | dunno | 21:10 |
JayF | I was taught, due to the old-days and variability in what "XX GB" disk sizes were, to always round off | 21:11 |
JayF | e.g. if I had a 512,342,000 byte disk, I'd make the raid partitions at 500,000,000 (whatever the aligned version of this would be) | 21:11 |
JayF | so I'm well on the side of ... leave more free space | 21:11 |
JayF | just bceause it prevents inconsistencies if you have disks with slightly different sizes | 21:11 |
JayF | I'm not sure best practices for managing raid arrays running with consumer hardware in 2005 apply here though LOL | 21:12 |
clarkb | TheJulia: I think the general appraoch makes sense but left a comment inline about specific implementation details I'm trying to reconcile with the commit message | 21:39 |
clarkb | I don't think opendev is using any lvm partitions in our images so impact to use either way is minimal, but maybe this deservers a functional test? If we aren't already building at least one of the booted test images with an lvm setup we should probably do so? | 21:39 |
TheJulia | clarkb: I started to stub out a 4k block test with ironic, since realistically we need to chain the whole thing together | 22:34 |
TheJulia | I'll likely get back to that tomorrow | 22:35 |
Generated by irclog2html.py 2.17.3 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!