fifofonix has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
jlebon has quit [Quit: leaving]
bgilbert_ is now known as bgilbert
<bgilbert>
dustymabe: started testing-devel. waiting for the config-bot sync before starting next-devel.
jlebon has joined #fedora-coreos
<bgilbert>
dustymabe: next-devel is running
jlebon has quit [Quit: leaving]
daMaestro has joined #fedora-coreos
ravanell_ has joined #fedora-coreos
ravanelli has quit [Ping timeout: 252 seconds]
gursewak has quit [Ping timeout: 246 seconds]
gursewak has joined #fedora-coreos
gursewak has quit [Ping timeout: 255 seconds]
ravanell_ has quit [Remote host closed the connection]
<bgilbert>
looks like the failures were: an amd64 iso-offline-install SIGKILL (presumed OOM), s390x grub-users failures that I understand, and ppc64le grub-users-fix failures that I don't.
<bgilbert>
the latter failed on both builds, and in both cases, the journal shows that the test actually passed (?)
baaash[m] has quit [Quit: You have been kicked for being idle]
paragan has quit [Ping timeout: 248 seconds]
paragan has joined #fedora-coreos
Arkanterian has joined #fedora-coreos
Arkanterian has quit [*.net *.split]
piwu7 has joined #fedora-coreos
piwu has quit [Ping timeout: 252 seconds]
piwu7 is now known as piwu
klaas_ has joined #fedora-coreos
klaas has quit [Ping timeout: 252 seconds]
crobinso has joined #fedora-coreos
jpn has quit [Ping timeout: 246 seconds]
jpn has joined #fedora-coreos
jpn has quit [Ping timeout: 255 seconds]
jpn has joined #fedora-coreos
nalind has joined #fedora-coreos
jpn has quit [Ping timeout: 248 seconds]
jpn has joined #fedora-coreos
plarsen has joined #fedora-coreos
<dustymabe>
lucab: looks like we need to run `testing` and `stable` again from scratch :(
<dustymabe>
that particular failure has been showing up in CI recently and we don't know why. It's intermittent. I asked some of the team to take a look at it earlier this week but no one has had a chance to figure it out yet. I'll open an issue for it here in an hour
piwu9 has joined #fedora-coreos
piwu has quit [Read error: Connection reset by peer]
piwu9 is now known as piwu
vgoyal has joined #fedora-coreos
<lucab>
dustymabe: ah, all the artifacts were properly pushed, I was hoping we could proceed from there anyway
jpn has quit [Ping timeout: 252 seconds]
jlebon has joined #fedora-coreos
jpn has joined #fedora-coreos
mheon has joined #fedora-coreos
<lucab>
dustymabe: but rerunning those, the new jobs finish immediately with "(no new build)". Should I force them?
jpn has quit [Ping timeout: 252 seconds]
<jlebon>
(reading from chat logs) lucab: yeah, you'll need to force it. marmijo is looking at that failure
<bgilbert>
well, if upstream changes the default, I wouldn't argue for turning it back off again
<jlebon>
that's not the same commitment as changing the default, but at least there's dogfooding going on even at the compiler level
<bgilbert>
+1
<walters>
For things where perf really matters, it's not just LTO, you also want PGO and even BOLT, neither of which we're doing (PGO just totally breaks in the koji model)
<walters>
For e.g. Firefox they absolutely need cross-language LTO even to make the Rust/C++ transition not terrible
<bgilbert>
without objection, I'd like to PR Afterburn and ssh-key-dir to drop the override
<lucab>
we have a blanket "LTO on" for the **release** profile almost everywhere, I'd say
<bgilbert>
my grep found those + rpm-ostree and zincati
<lucab>
Colin Walters: ostree-ext is currently on thin-lto, do you want to turn that off too?
<walters>
thin lto works fine with incremental builds
<walters>
I am only battling against `lto = true` by default
<walters>
the difference between 1 second and 1 minute is palpable
<lucab>
ack, let's stick to only killing "fat" LTO everywhere then
<dustymabe>
jlebon: regarding https://github.com/coreos/rpm-ostree/pull/4122 - do we need to consider silverblue and IoT here? i.e. should we try to get that fix promoted into the GA of 37 (meaning FE or Blocker bug)
<dustymabe>
lucab: that s390x stable build is getting close to being done
<walters>
(I upgraded my workstation to a new i9-13900k, it absolutely plows through compile jobs; building ostree-ext from scratch takes 23 seconds on this, over 4 minutes on my older i7-8665U laptop...so I now definitely notice a full minute for LTO)
<bgilbert>
lucab walters: thanks for bringing this up
<lucab>
dustymabe: yay, that's the last straggler. I've pre-check and ticked all the other boxes, we can jump to release jobs after that.
<dustymabe>
lucab: I think so
<jlebon>
dustymabe: i'm not sure if it's blocker-level. the policy still does get updated. it just "dirties" system state unnecessarily
<dustymabe>
jlebon +1
<jlebon>
i also think there's a path towards salvaging those machines so they're back on the canonical policy
poppajarv1 has joined #fedora-coreos
* dustymabe
grabbing food
poppajarv has quit [Ping timeout: 272 seconds]
poppajarv1 is now known as poppajarv
bytehackr has quit [Ping timeout: 255 seconds]
<lucab>
triple releases almost done, what time do we want to start the rollouts? 20:00 UTC (~2h from now)?
jpn has quit [Ping timeout: 246 seconds]
strigazi has quit [Ping timeout: 252 seconds]
strigazi has joined #fedora-coreos
aaradhak has joined #fedora-coreos
<bgilbert>
lucab: sgtm
<lucab>
nope, oscontainer pushing broke for testing
<dustymabe>
wow SIGKILL 9 (which is usually what we see when the job runs out of memory)
<dustymabe>
lucab: should be safe to re-run it (though we can see if stable goes through fine first)
<dustymabe>
s/stable/next
<lucab>
stable is already done, I started the jobs staggered
<lucab>
ah yes, next is ongoing right now
<dustymabe>
yeah feel free to go ahead and re-run - the job is pretty well idempotent
<dustymabe>
at least the parts that have already run are
HappyMan has quit [Read error: Software caused connection abort]