I upgraded one of our servers to 12.1 earlier this month, and ever since then I've noticed a disturbing trend in CPU usage:
When investigating this, I noticed it was basically all due to newsyslog gobbling up CPU time, and seemingly more and more as time went on. I tried killing all the running processes (where you see a dip on the 27th), hoping it was just that one was hung up causing others to hang as well, but when they restart from cron, the problem continues, as you can see.
I've run
...and then appears to just hang idefinitely. net-snmp.conf by default tries to rotate /var/log/snmpd.log, but as that doesn't exist on my system, I've commented it out, so the file is currently nothing but comments. Any idea why it hangs and then gobbles up CPU, or where I should go in terms of debugging from here?
When investigating this, I noticed it was basically all due to newsyslog gobbling up CPU time, and seemingly more and more as time went on. I tried killing all the running processes (where you see a dip on the 27th), hoping it was just that one was hung up causing others to hang as well, but when they restart from cron, the problem continues, as you can see.
I've run
newsyslog -nv
, which shows the following:
Code:
Processing /etc/newsyslog.conf
Found: <include> /etc/newsyslog.conf.d/*
Found: <include> /usr/local/etc/newsyslog.conf.d/*
Processing /etc/newsyslog.conf.d/ftp.conf
Processing /etc/newsyslog.conf.d/opensm.conf
Processing /etc/newsyslog.conf.d/pf.conf
Processing /etc/newsyslog.conf.d/sendmail.conf
Processing /usr/local/etc/newsyslog.conf.d/net-snmp.conf