Created attachment 867723 [details] The patch to fix the issue. Description of problem: using snmp monitor cpu usage with more than 100 processors. hrProcessorLoad will returns incorrect values for processor #'s > 100 Version-Release number of selected component (if applicable): RHEL 5.x RHEL 6.x How reproducible: Steps to Reproduce: 1. using snmp monitor cpu usage with more than 100 processors. Actual results: hrProcessorLoad will return incorrect values for processor #'s Expected results: hrProcessorLoad will return correct values for processor #'s (letter than 100) Additional info: http://mailing.unix.net-snmp-users.narkive.com/o21D50GY/hrprocessorload-returns-incorrect-values-for-processor-s-100 The file agent/mibgroup/hardware/cpu/cpu_linux.c: netsnmp_cpu_arch_load() use below hard code to skip cpuN b1 = b2+5; /* Skip "cpuN " */ When processor more than 100, the value will be wrong.
The patch have be applied to net-snmp upstream: http://git.code.sf.net/p/net-snmp/code: commit eef2f64b46357b353c79504bc593535ebe7421e7 Author: Niels Baggesen <nba.net> Date: Thu Jan 23 16:27:07 2014 +0100 cpu_linux: support systems with more than 100 cpus.
Thanks for the report. RHEL 5 is in maintenance mode and only critical bugs are fixed. Please contact your Red Hat support, e.g. at https://access.redhat.com/ and negotiate an exception. I've cloned the bug into newer RHEL releases (6+7) and it should be better there. However, a support ticket would speed up the bugfix there.