[KUDU-1290] CHECK failures when some CPUs are disabled

• Type: Bug
• Status: Resolved
• Priority: Major
• Resolution: Fixed
• Affects Version/s: 0.6.0
• Fix Version/s: 0.7.0
• Component/s: util
• Target Version/s: 0.7.0
• Code Review: http://gerrit.cloudera.org:8080/#/c/1864/
• Assignee: Todd Lipcon
• Reporter: Todd Lipcon
• Created: 07/Jan/16 16:27
• Updated: 22/Jan/16 04:47
• Resolved: 22/Jan/16 04:47

Description

"Bsinghal" on the Cloudera forum reported this bug:

I have a server with 48 cores. I disabled 24 cores using the following script
for((i=12;i<=23;i++)); do echo 0 | sudo tee /sys/devices/system/cpu/cpu$i/online; done
for((i=36;i<=47;i++)); do echo 0 | sudo tee /sys/devices/system/cpu/cpu$i/online; done

After this when I start kudu-master service, it fails with the following error:
F0107 16:27:30.428330 4140 locks.h:175] Check failed: cpu < n_cpus_ (31 vs. 24)

/proc/cpuinfo shows that only 24 processors are there, but somehow kudu-master thinks there are 31 processor.