I've been troubleshooting an issue with transfer speeds on my servers, and one of the steps was to try and upgrade my main server NODE, to RC2 to match VOID.
After doing so, I've noticed my CPU (Intel i7-4770) is way busier than it was on 6.8.3.
On 6.8.3, my plex docker could support 8-10 streams. On RC2, my Plex server is choking to death trying to transcode a single episode while I'm the only person online. I get stuttering/buffering and the transcoder is only reporting a 0.8 conversation rate (meaning plex can't transcode fast enough to keep up). If I have more than one person using it, forget about it, its god awful for everyone who tries to use it.
I saw some posts (here, here) in the RC2 release thread saying people were experiencing similar issues with their VMs. While I don't run any VMs I do have about 8 dockers and they all feel much much slower and regularly 100% all available cores even for what I would consider light duty tasks.
Downgrading to 6.8.3 resolves my performance issues on NODE, plex can keep up with multiple streams and my overall CPU usage over a given time is halved. This is with NO OTHER CHANGES to the system.
Steps to reproduce:
-upgrade to 6.9.0-RC2 and watch your CPU performance tank.
Downgrade to 6.8.3 and it all returns to "normal".
node-diagnostics-20210204-0645.zip
EDIT: Just found ljm42's post here about the cpu governor in tweaks and tips, mine was indeed on power saver mode so I will test and see if this resolves my performance issues on RC2.
EDIT2: Yeah that seems to have done it. I never would have guessed the CPU governor setting got changed, so I'm glad I found that post.
Recommended Comments
Join the conversation
You can post now and register later. If you have an account, sign in now to post with your account.
Note: Your post will require moderator approval before it will be visible.