On the master branch:
This PR branch:
Also consider “CPU Usage” charts provided by CI.
Overlooked in cda62657e95a90a5fd61ba43e2acbd407e3a4135 (bitcoin/bitcoin#25929).
183@@ -184,7 +184,7 @@ task:
184 - netsh int ipv4 set dynamicport tcp start=1025 num=64511
185 - netsh int ipv6 set dynamicport tcp start=1025 num=64511
186 # Exclude feature_dbcrash for now due to timeout
187- - python test\functional\test_runner.py --nocleanup --ci --quiet --combinedlogslen=4000 --jobs=4 --timeout-factor=8 --extended --exclude feature_dbcrash
188+ - python test\functional\test_runner.py --nocleanup --ci --quiet --combinedlogslen=4000 --jobs=6 --timeout-factor=8 --extended --exclude feature_dbcrash
0 - python test\functional\test_runner.py --nocleanup --ci --quiet --combinedlogslen=4000 --jobs=12 --timeout-factor=8 --extended --exclude feature_dbcrash
There should be no risk in saturating the CPU, given enough memory
Updated 6fbd173d8a4519967d75d2707b2285d62faa4424 -> dc869c6a77c4542d4d9dcefdec577d1876e51061 (pr26297.01 -> pr26297.02, diff):
Let CI task finish to see the actual effect of a possible over-subscription.
Looks like this makes the intermittent issues more common
Indeed. Reverted back to 6fbd173d8a4519967d75d2707b2285d62faa4424 (pr26297.01), which was already ACKed by @aureleoules.
The following sections might be updated with supplementary metadata relevant to reviewers and maintainers.
Reviewers, this pull request conflicts with the following ones:
If you consider this pull request important, please also help to review the conflicting pull requests. Ideally, start with the one that should be merged first.
ACK 6fbd173d8a4519967d75d2707b2285d62faa4424
Lgtm
ACK 6fbd173d8a4519967d75d2707b2285d62faa4424
Makes sense to increase number of jobs equal to number of CPUs.
The merge commit description included hebasto’s comment even tough it’s not a ACK @MarcoFalke, probably not that big of a deal though.
Edit: Sorry for the ping I confused DrahtBot and the github-merge script :grimacing:
ACKs for top commit: hebasto: Indeed. Reverted back to https://github.com/bitcoin/bitcoin/commit/6fbd173d8a4519967d75d2707b2285d62faa4424 (pr26297.01), which was already ACKed by @aureleoules. aureleoules: ACK https://github.com/bitcoin/bitcoin/commit/6fbd173d8a4519967d75d2707b2285d62faa4424 jarolrod: ACK https://github.com/bitcoin/bitcoin/commit/6fbd173d8a4519967d75d2707b2285d62faa4424 shaavan: ACK https://github.com/bitcoin/bitcoin/commit/6fbd173d8a4519967d75d2707b2285d62faa4424
Why
So that intermittent issues are happening more consistently and can then hopefully be found and fixed before they make it into the master branch