qschulz has quit [Remote host closed the connection]
qschulz has joined #yocto
vvn has quit [Quit: WeeChat 4.5.1]
davidinux has quit [Ping timeout: 260 seconds]
davidinux has joined #yocto
jclsn has quit [Ping timeout: 272 seconds]
jclsn has joined #yocto
frgo has joined #yocto
vvn has joined #yocto
chep has quit [Ping timeout: 276 seconds]
chep has joined #yocto
tammranil has quit [Remote host closed the connection]
tammranil has joined #yocto
jmd has joined #yocto
frgo has quit [Remote host closed the connection]
frgo has joined #yocto
rob_w has joined #yocto
frgo has quit [Ping timeout: 252 seconds]
frgo has joined #yocto
frgo has quit [Remote host closed the connection]
ndeuteron has joined #yocto
<ndeuteron>
Hi. What causes yocto to parse recipes every time I run it, rather than caching? I thought it was BB_DONT_CACHE but I've set that to 0 and still no joy.
rfuentess has quit [Read error: Connection reset by peer]
rfuentess has joined #yocto
wojci has joined #yocto
grma has joined #yocto
druppy has joined #yocto
frgo has joined #yocto
druppy has quit [Ping timeout: 246 seconds]
mckoan|away is now known as mckoan
<LetoThe2nd>
ndeuteron: I would assume any kind of modification in a file.
<RP>
ndeuteron: try not setting that at all...
leon-anavi has joined #yocto
frgo has quit [Read error: Connection reset by peer]
frgo_ has joined #yocto
frgo_ has quit [Remote host closed the connection]
frgo has joined #yocto
Harish has joined #yocto
<ndeuteron>
LetoThe2nd: It does it with no changes to any files. Back to back builds with nothing changing in between.
<ndeuteron>
RP: Hmm, will do.
makurz has joined #yocto
makurz has quit [Client Quit]
<ndeuteron>
RP: That works. Thanks!
ablu has quit [Ping timeout: 276 seconds]
ablu has joined #yocto
<ndeuteron>
Is that a bug? Confusing at the very least.
florian has joined #yocto
Kubu_work has joined #yocto
florian_kc has joined #yocto
rfuentess has quit [Remote host closed the connection]
rfuentess has joined #yocto
rfuentess has quit [Read error: Connection reset by peer]
rfuentess has joined #yocto
prabhakalad has quit [Ping timeout: 245 seconds]
prabhakalad has joined #yocto
Habbie has quit [Ping timeout: 272 seconds]
ThomasRoos has joined #yocto
frgo has quit [Remote host closed the connection]
berton has joined #yocto
<RP>
ndeuteron: is it a bug? yes and no. Many of our values don't accept "0" as disable :/
<RP>
it requires a lot more processing than "is this set or not?"
frgo has joined #yocto
halloy9196 has joined #yocto
<halloy9196>
I'm annoyed by the do_rootfs task. switching from a single nvme ssd, to a raid0 dual nvme. can i expect any significant gain in duration reduction?
<halloy9196>
Watching IO during this step does not look promising, but this task consists mostly of copying, to I'd expect an impact
halloy9196 is now known as jarvis-owl
aardo has joined #yocto
<mcfrisk>
jarvis-owl: tune build machine kernel vm to cache io operations in RAM, rm_work or manual cleanup will remove the files anyway. This does not help if package manager, which works in series not parallel, calls sync()/fsync() all the time. works for all recipe compilations, not just rootfs. and scales better than building on tmpfs will can run out of space without overflowing to real disk.
<mcfrisk>
to get rid of sync/fsync calls, eatmydata wrapper
ardo has quit [Ping timeout: 244 seconds]
<jarvis-owl>
nice, thanks for the hints! i'm going to test that :)
<RP>
mcfrisk: if the package managers run under pseudo, we intercept the sync/fsync calls FWIW
<AdrianF>
ThomasRoos: Thank you for raising the issue of address sanitizers. We have just decided to do the same. Would it be possible to open a bug that summarizes the status of your work and findings and to which we could also contribute?
<ThomasRoos>
AdrianF Hi Adrian, basically do this: CXXFLAGS:append:aarch64 = " -DSANITIZER_CAN_USE_ALLOCATOR64=0" in gcc-sanitizers.inc work around it for arm64 (tested with a recipe containing an error)
<ThomasRoos>
I was about to send a patch, but not sure what is the correct mailing list. rburton
<ThomasRoos>
Also tested it with arm32 - this is working without the patch. So this should be good.
<mcfrisk>
RP: good to hear, I think I saw some call paths in the past which allowed native sync/fsync calls to pass. Could have been in normal build tasks, not do_rootfs. These are easy to spot once the vm config is applied. IO going to disk should stop as long as RAM is available. https://pcp.io or similar perf data loggers can show CPU, memory, IO and networking load when builds run.
<mcfrisk>
(default 5 second wait for ext3/4 to do background writing is really bad on build machines)
<RP>
mcfrisk: it would be interesting to see where the problem calls come from. I know we do sync the qemu images to disk before testimage because of the opposite problem - IO storm during qemu causing test failures
<mcfrisk>
my builds on aarch64 machine are fast enough already so had no reason to dig deeper :)
<RP>
mcfrisk: right, NVMe is great too
<mcfrisk>
/proc/sys/vm/dirty_background_ratio at 10, and /proc/sys/vm/dirty_writeback_centisecs at 500 on my current machine. IO writes are not bugging me though they can kill the device at some point. rm_work is enabled so writes should actually fit to RAM all the time when doing mostly upstream image builds.
<halstead>
I've scaled the wiki up to handle increased load while dealing with an intense data harvesting event. It should be back online now.
mulk has quit [Ping timeout: 248 seconds]
ptsneves has quit [Ping timeout: 245 seconds]
<mischief>
halstead: is that one of those ai crawlers? :)
<halstead>
mischief: my hunch is yes. It's built to look like a bunch of Macbooks are all browsing the wiki history from all over the world at once though.
mulk has joined #yocto
Xagen has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
Kubu_work has joined #yocto
<mischief>
claudebot took down my personal git server so i had to robots.txt it
khem has joined #yocto
wojci has joined #yocto
goliath has joined #yocto
ehussain has joined #yocto
dankm has quit [Remote host closed the connection]
dankm has joined #yocto
florian_kc is now known as florian
ehussain has quit [Remote host closed the connection]
sakoman has quit [Ping timeout: 272 seconds]
nerdboy has quit [Ping timeout: 252 seconds]
sakoman has joined #yocto
miau_miau has quit [Quit: Client closed]
nerdboy has joined #yocto
nerdboy has quit [Changing host]
nerdboy has joined #yocto
jmd has quit [Remote host closed the connection]
tangofoxtrot has quit [Remote host closed the connection]
<rburton>
halstead: <shakes fist>
<rburton>
the sooner a few big orgs get together and have a crowd-sourced live IP block list the better
tangofoxtrot has joined #yocto
savolla has quit [Quit: WeeChat 4.4.3]
druppy has joined #yocto
cyxae has quit [Quit: cyxae]
wojci has quit [Ping timeout: 245 seconds]
zhmylove has quit [Remote host closed the connection]
druppy has quit [Ping timeout: 244 seconds]
Kubu_work has quit [Quit: Leaving.]
sotaoverride is now known as Guest9588
Guest9588 has quit [Killed (silver.libera.chat (Nickname regained by services))]