[08:09] <laney> jawn-smith: it would be worth filing a bug/card to fix that flakiness, especially given ubuntu-image is something we maintain ...
[09:32] <xnox> seb128:  was on pto, will be submitting it today. (initramfs-tools zstd bionic sru)
[09:32] <seb128> xnox, thanks!
[09:32] <xnox> sil2100:  preinstalled images are here =) nice. But also wondering why arm64 one is 240MB bigger than amd64 one!
[11:11] <sil2100> xnox: hey! Good to hear! I didn't even have time to take a look at those yet ;)
[13:36] <jawn-smith> laney: as of 21.10 ubuntu-image will be completely re-written in golang, and one of my goals during the rewrite will be tests that aren't flakey
[18:15] <xnox> seb128:  heya, i am confused.
[18:15] <xnox> bionic's unmkinitramfs does not have lz4 support either. and we have been building all our isos with lz4 initrds for a few cycles now.
[18:15] <xnox> are you sure that the system that checks things is running stock initramfs-tools from bionic.... and doesn't already have some local backports installed?
[18:16] <xnox> i guess i could add both lz4 and zstd support to unmkinitramfs in bionic, but just checking to be sure. In case we do bionic sru and that doesn't actually fix our iso static CI
[20:50] <seb128> xnox, hum, I need to check, it's not impossible that someone just hacked the script on the CI system to workaround the issue rather than getting in fixed in a SRU
[21:00] <seb128> xnox, so yeah, the machine has a local version of unmkinitramfs which was probably created to workaround tghe missing lz support, I wil try to add zstd locally, would still be useful to have an official SRU, it's not the first time I have to do local hacks due to the missing formats