seninha has quit [Remote host closed the connection]
<mischief>
can sharing sstate dirs between different bitbake processes race?
starblue has quit [Ping timeout: 255 seconds]
starblue has joined #yocto
dmoseley has quit [Ping timeout: 264 seconds]
xmn has joined #yocto
Vonter has joined #yocto
<mischief>
also for some reason do_bundle_initramfs is recompiling my whole kernel (single threaded because it doesnt pass PARALLEL_MAKE) ..
paulg has quit [Ping timeout: 265 seconds]
xmn has quit [Ping timeout: 265 seconds]
beneth has quit [Remote host closed the connection]
AKN has joined #yocto
alessioigor has joined #yocto
alessioigor has quit [Client Quit]
nemik has quit [Ping timeout: 250 seconds]
nemik has joined #yocto
nemik has quit [Ping timeout: 252 seconds]
nemik has joined #yocto
dtometzki has joined #yocto
<RP>
Ch^W_: we can probably do that
<RP>
mischief: sstate shouldn't race
<mischief>
RP: hm. we see some build failures sometimes for do_package_write_deb_setscene
<mischief>
er.. maybe red herring in the log. we also see do_package_qa_setscene: fetcher failure: unable to find file file:// ....
<mischief>
which seems like an issue with sstate. perhaps sstate is just corrupt and should be wiped?
<RP>
mischief: can you share an example log of the whole failure?
<RP>
mischief: I can't comment based on that information
<mischief>
is there anything in particular that would be helpful? there is quite a bit of internal code in our build.
<RP>
mischief: for setscene to fail the "pre" task check due task init must have said it exists, then at setscene time it didn't. Are you cleaning things from sstate?
<RP>
mischief: was both the siginfo and the sstate file missing or just one of them?
<RP>
there was recently a race reported on the mailing list using SSTATE_DIR on some kind of kubernetes setup with a weirder filesystem that didn't support hardlinks
<mischief>
RP: we don't prune sstate automatically.. probably should since i've had to clean it manually due to a full disk.
<RP>
mischief: are builds running when you pruned it?
<mischief>
no, haven't pruned it in months
<mischief>
last time i pruned i just ran a find -delete command with an atime filter
<mischief>
it looks like the do_package_qa_setscene error was on a package_qa.tgz.siginfo file, but the package_qa.tgz seems to be present
<mischief>
maybe the result of my previous pruning gone wrong
nemik has quit [Ping timeout: 265 seconds]
nemik has joined #yocto
nemik has quit [Ping timeout: 265 seconds]
nemik has joined #yocto
<mischief>
afaik its safe to delete the sstate so i might just wipe our sstate and sstate mirror and eat the rebuild cost, since our daily builds seems to hit this kind of thing often
lexano has quit [Ping timeout: 252 seconds]
lexano has joined #yocto
amitk has quit [Quit: leaving]
amitk has joined #yocto
yudjinn[m] has quit [Quit: You have been kicked for being idle]
amitk_ has joined #yocto
amitk has quit [Ping timeout: 265 seconds]
nemik has quit [Ping timeout: 252 seconds]
nemik has joined #yocto
florian_kc has joined #yocto
nemik has quit [Ping timeout: 265 seconds]
nemik has joined #yocto
<RP>
mischief: none of the CI runs the cleansstate task?
<RP>
mischief: removing it all and rebuilding should be safe