SOLVED: High CPU temp after upgrade from 22.12.3.3

[fireblade70]

I unintentionally upgraded from 22.12.3.3 (from a Home Assistant integration but that's another story) and straight after without any changes my CPU power consumption almost doubled (have wattage meter but also see in image). Only have two VM's (HAOS & Windows) running and 1 app (plex). I have nothing special in VM configs (min recommended) Any idea how to resolve? Prior to upgrade, the Avg was about the same and all threads evened out but now the erratic usage of the threads is heating up the CPU & sucking electric.

MB: GigaByte B760M_DS3H
CPU: i5-13500
RAM: 64Gb (50% utilised)
Storage 2x WD HUS728T8TALE6L4 8TB, 1x Samsung SSD 840 EVO 250Gb, 1x WD Blue SN570 500GB, 1x Samsung SSD 970 EVO 500GB (No Controller)
Network: MB interface

 

[morganL]

I unintentionally upgraded from 22.12.3.3 (from a Home Assistant integration but that's another story) and straight after without any changes my CPU power consumption almost doubled (have wattage meter but also see in image). Only have two VM's (HAOS & Windows) running and 1 app (plex). I have nothing special in VM configs (min recommended) Any idea how to resolve? Prior to upgrade, the Avg was about the same and all threads evened out but now the erratic usage of the threads is heating up the CPU & sucking electric.
View attachment 69440

MB: GigaByte B760M_DS3H
CPU: i5-13500
RAM: 64Gb (50% utilised)
Storage 2x WD HUS728T8TALE6L4 8TB, 1x Samsung SSD 840 EVO 250Gb, 1x WD Blue SN570 500GB, 1x Samsung SSD 970 EVO 500GB (No Controller)
Network: MB interface

Could you also share the wattage readings...

Usage = 4% is quite low.

The instrumentation within Cobia has been updated and hence it might just give different readings even if it's similar.

 

[fireblade70]

Could you also share the wattage readings...

Usage = 4% is quite low.

The instrumentation within Cobia has been updated and hence it might just give different readings even if it's similar.

Usage bounces around so much and hence temp cores and wattage almost doubled. The following was running same VM's(load) with update carried out just before 4pm (ignore straight line) I shut HA off. In the meantime I tried creating 2 new identical VM's with same zvol's and still same result.

 

[morganL]

Usage bounces around so much and hence temp cores and wattage almost doubled. The following was running same VM's(load) with update carried out just before 4pm (ignore straight line) I shut HA off. In the meantime I tried creating 2 new identical VM's with same zvol's and still same result. View attachment 69491

So you think power doubles from 100W to 200W?

If you identify the load that causes this excess power consumption, we should report a bug.

 

[fireblade70]

So you think power doubles from 100W to 200W?

If you identify the load that causes this excess power consumption, we should report a bug.

Yes. I isolated it down to just running one my Home Assistant VM and CPU spikes are still erratic despite HA reporting 10-16% cpu usage. I've included image from a top command and I have suspicions the asycio_loop is taking up alot and also bouncing around 60-70%! Has this anything to do with THIS or because this processor has a hybrid architecture (although it's been absolutely fine on 22.12.3.3)?

 

[fireblade70]

Yes. I isolated it down to just running one my Home Assistant VM and CPU spikes are still erratic despite HA reporting 10-16% cpu usage. I've included image from a top command and I have suspicions the asycio_loop is taking up alot and also bouncing around 60-70%! Has this anything to do with THIS or because this processor has a hybrid architecture (although it's been absolutely fine on 22.12.3.3)?

Sorry for got image

 

[fireblade70]

Could you also share the wattage readings...

Usage = 4% is quite low.

The instrumentation within Cobia has been updated and hence it might just give different readings even if it's similar.

I think I may of resolved it! I suspect it was a Truenas API - Home Assistant Truenas Integration (same one which automated the update of my Truenas!!). From 210k errors in logs in HA, filling database. So removed it and it's all settled down! It's a shame Truenas Scale doesn't have GUI feature to show cpu usage by VM or App. I will check over next 24hrs if settled and update.

 

[morganL]

I think I may of resolved it! I suspect it was a Truenas API - Home Assistant Truenas Integration (same one which automated the update of my Truenas!!). From 210k errors in logs in HA, filling database. So removed it and it's all settled down! It's a shame Truenas Scale doesn't have GUI feature to show cpu usage by VM or App. I will check over next 24hrs if settled and update.

Feel free to make a feature request in Jira... use your case as an example.

Glad you resolved.