orivej has quit [Quit: No Ping reply in 180 seconds.]
orivej has joined #nixos-dev
init_6 has joined #nixos-dev
genesis has quit [Remote host closed the connection]
genesis has joined #nixos-dev
jtojnar has joined #nixos-dev
<gchristensen>
srhb, andi- 'round?
<andi->
gchristensen: yes, with bad train wifi but here
<gchristensen>
proposed changes to job / core count allocations & I'd like feedback
<gchristensen>
we got a new aarch64 builder (packet-t2a5-qc-centriq-1, 46 cores @ 2.5ghz, 96g ram) with this new capacity, I propose we cut the # of jobs on t2a-3 from 80 to 40, and give each job 2 cores, and set on c2-centriq-1 to 20 jobs, and give each job 2 cores
<andi->
sounds reasonable, give a bit of headroom for builds that do consume more ressources. ARe those 46 cores split across multiple sockets or in one?
<gchristensen>
all in one socket
<gchristensen>
a note on the current split for t2a-3 is it has 96 cores, running 80 jobs with 1 core each
* andi-
tries to access the grafana
<gchristensen>
t2a-2 is 96 cores too, and accepts 2 jobs, each with 45 cores and it only accepts big-parallel jobs
<andi->
sounds okay to me. Some more flexible scheduling would be nice to use the spare capacity on t2a-2 but well..
<gchristensen>
yea
orivej has quit [Ping timeout: 245 seconds]
init_6 has quit [Ping timeout: 250 seconds]
<srhb>
gchristensen: Seems reasonable to me!
<gchristensen>
+1 thanks!
orivej has joined #nixos-dev
phreedom has quit [Remote host closed the connection]
phreedom has joined #nixos-dev
Synthetica has joined #nixos-dev
q6AA4FD has joined #nixos-dev
lassulus has quit [Ping timeout: 268 seconds]
lassulus_ has joined #nixos-dev
lassulus_ is now known as lassulus
orivej has quit [Ping timeout: 240 seconds]
orivej has joined #nixos-dev
lassulus_ has joined #nixos-dev
lassulus has quit [Ping timeout: 268 seconds]
lassulus_ is now known as lassulus
orivej has quit [Ping timeout: 272 seconds]
MichaelRaskin has joined #nixos-dev
<jtojnar>
samueldr: I do not really know much about ELF, ABIs and stuff like that so I am pretty much done now
<samueldr>
jtojnar: you probably went further than I would, good sleuthing, but hm
<samueldr>
just had a thought
<samueldr>
ah no, you tried it, matching the other libraries, e.g. same libpng
<jtojnar>
assuming I did not make some mistake
<jtojnar>
but yeah, nix-store --query --tree result/ seems to confirm the same libraries are used
<samueldr>
what do you think should be done? like the Qt situation it's possible this may become a well known issue of mixing stable+unstable (though hopefully this one is limited to this upgrade?)
<jtojnar>
well we could require that all gtk applications use wrapGAppsHook
<jtojnar>
including GTK2
<samueldr>
that may require the whole of Qt to do it too, due to how qtbase links to gdk_pixbuf (iirc) I guess for their gnome compatibility?
<samueldr>
the list I linked to in the issue excluded all Qt things
<samueldr>
(though I haven't seen this issue crop up in Qt apps, but neither did I try to find one)