Actually no. I can't see any spikes in CPU usage when this happens. When /lib/systemd/system/lxcfs.service was running with the -l flag, there was a problem with one tool. The CPU usage was always at 100%, and if I recall, correctly, it was htop when top, glances and monitoring solution showed...
Some time ago I started to receive syslog errors like this every few seconds for all containers in a physical node:
bindings.c: 4526: cpuview_proc_stat: cpuX from /lxc/CID/ns has unexpected cpu time: 9164403 in /proc/stat, 9393872 in cpuacct.usage_all; unable to determine idle time
I added flag...
This site uses cookies to help personalise content, tailor your experience and to keep you logged in if you register.
By continuing to use this site, you are consenting to our use of cookies.