On 06/04/2015 08:32 AM, George Dunlap wrote:
> for i in {1..80} ; do xl create null.cfg name=\"t$i\"
> kernel=\"/boot/vmlinuz-3.18.12-11.el7.x86_64\" memory=\"40\"
> on_crash=\"preserve\" ; done
>
> And then used the shell snippet from the bug report to generate load:
>
> while true; do xl list &> /dev/null; usleep 5000;done
>
> With xenstored normally the CPU load was about 30%. Mounted under
> tmpfs, it was slightly smaller, around 20%, but swung around wildly.
> With oxenstored it was about 15%.
>
> This is with xfs as the filesystem.
>
> That doesn't seem big enough to warrant a lot of effort: if you can
> reproduce the 75% numbers from the bug report I'll pursue it further,
> otherwise I'll probably just leave it.
I'm actually more concerned about I/O wait than CPU usage. When shutting down or starting a lot of guests the wait time could be significantly increased.
I only started to look into this because I was doing load testing and it was taking much longer to shut everything down then it seemed like it should.
Running multiple versions of xl in parallel (especially xl destroy) can make many, many backups of tdb such that there's hundreds of MB of backup copies.
--Sarah