Python CPU Spikes

[Dinez]

Hello guys,

I've noticed that i get heavy CPU spikes that reach up to 60% of CPU usage.

If anybody could help out resolve this issue that would be great.

My System:
NZXT N7 Z390
Intel(R) Core(TM) i7-9700F CPU @ 3.00GHz
32 Gigs of ram
Kingston NV2 500GB M.2 NVMe
Powertech 1 Gig NIC (not using currently)

 

[Ericloewe]

What, exactly, is there to resolve? The system does work, consuming CPU time.

 

[Davvo]

Are you asking us to identify where that python spawn is coming from without any context? That's a tad hard.

 

[Dinez]

What, exactly, is there to resolve? The system does work, consuming CPU time.

So this is normal then?

 

[Dinez]

Are you asking us to identify where that python spawn is coming from without any context? That's a tad hard.

Dont really have any context due to the fact that i dont know where its coming from or what the next steps is to dig further into.

 

[winnielinnie]

Perhaps it's from the main middlewared process?

When you have htop open, press F5 to view a "tree" layout.

 

[winnielinnie]

The system does work, consuming CPU time.

That is a good deal of CPU, though. Even for SCALE, surely?

On my Core system, with qBittorrent running in a jail 24/7, my CPU remains at 0 - 2%.
(My CPU graph looks like a patient flatlined on a hospital bed.)

 

[anodos]

That looks like the ZFS worker pool. Do you have some monitoring solution that's hitting our API endpoints? For example, I seem to recall a home assistant plugin that polls some rather expensive ZFS-related API endpoints.

 

[Dinez]

Perhaps it's from the main middlewared process?

When you have htop open, press F5 to view a "tree" layout.

Yeah seems like middleward. What could i do to dig further into it?

 

[Dinez]

That looks like the ZFS worker pool. Do you have some monitoring solution that's hitting our API endpoints? For example, I seem to recall a home assistant plugin that polls some rather expensive ZFS-related API endpoints.

Deleted all snapshots and removed all periodic Snapshots tasks. If thats what you ment.

 

[anodos]

Deleted all snapshots and removed all periodic Snapshots tasks. If thats what you ment.

I'm talking about applications that poll our ZFS / storage related endpoints (for instance running apps or external). There are some that do this. That said, CPU utilization isn't necessarily surprising, there are periodic background tasks related to system / pool monitoring.

 

[Dinez]

Semi-Solution: If you have a ZFS pool that has been degraded you get those weird cpu spikes.
On my end a cable was loose on one of my hdd's and thats why i had those spikes.
After i put that cable back in everything works kinda fine now. The spikes are less frequent.