Jail CPU loads in Main FreeNAS GUI graph

[Brad1976]

Hey all!

I checked the Docs and can't seem to find my answer... I have Plex running in an iocage jail and it does some transcoding from time to time. When I check the dashboard within the Plex GUI, I can see the CPU load sometimes goes as high as 25%, so I know it's doing something and CPU is definitely in use...

However, when I look at the graph from the OS main GUI, my CPU load is pegged right down at the bottom and it's a flat line... (And no I'm not being punny and saying the system is dead, it's very much alive...) lol

Now, I am very understanding about the segregation of jail space and system space with regard to system resources. Thus, I am pretty sure that the graph on the main GUI is only showing what the OS itself is using, and the graph within Plex is showing the usage for the cores that it's monopolizing...

1) Primarily I'm wondering if I can get a quick confirmation if I'm correct in my suspicion that they're separate... I'm currently on 11.2-U6, and hoping to get the time to move up to U7 by the end of this week.

2) If I was to want to get something to referance the CPU usage as a holistic view of all my combined oS and all jails in one graph, how does one go about that; if it's even possible? (Totally expecting to be told that I will need to move from jails to vm's here... If that is the case... Please also with me good luck to not get divorced again as I dive into another learning curve!!! At my age, my brain is so full that in order to learn something new I'll have to forget how to do something else like brushing my teeth, or how to poop!!!!!!!!!) Lol

Thanks in advance and hope all is well for all of you!!!

 

[sretalla]

Primarily I'm wondering if I can get a quick confirmation if I'm correct in my suspicion that they're separate

No. The charts show overall CPU including jail.

If I was to want to get something to referance the CPU usage as a holistic view of all my combined oS and all jails in one graph, how does one go about that

That's what you have in the graphs.

Maybe there's something wrong with your graphs...

 

[Brad1976]

Thanks sretalla, but OH CRAPOLA!!! Truthfully that is NOT what I was expecting to hear. Now I'm going to have to do some troubleshooting, however I'll first install the U7 update and report back...

I know this isn't a FreeNAS question, but in order to troubleshoot this in U7, I think I will have to select a bunch of items and attempt to Optimize them to create a cpu load... Anyone out there have any other suggestions on how to temporarily create a substantial cpu load to see if I can create a bump in my graph?

 

[Fredda]

Anyone out there have any other suggestions on how to temporarily create a substantial cpu load to see if I can create a bump in my graph?

You can simply run an empty loop in the shell, this should use up one CPU.
In bash you can simply run this one liner: while true ; do sleep 0 ; done
In csh you can do it that way:

Code:

~# while (1)
while? sleep 0
while? end

You should also check if it make a difference, if you run it in a jail or in FreeNAS directly.

 

[sretalla]

Also:
Storage Drives: 8 x 6Tb Segate Ironwolf RaidZ1 in 2 x 4 drive vdevs

This is risky and not a good configuration in the eyes of the forum.

Unless you have a specific need for the additional IOs of the second VDEV, you would protect your data much better with an 8-wide single RAIDZ2 VDEV.

As it stands, any single disk failure leaves you with no redundancy and at imminent risk of complete pool loss.

The resilver time for 6TB makes it more likely that you could see such an event during the failed disk replacement.

If you're OK with that, no problem, but you may want to give it come consideration.

 

[sretalla]

Anyone out there have any other suggestions on how to temporarily create a substantial cpu load to see if I can create a bump in my graph?

There is a freebsd utility called stress... https://www.freebsd.org/cgi/man.cgi?stress(1)

It isn't part of FreeNAS, so you should install it in a jail only, but you can then run it either in the jail or from the host using /mnt/pool/iocage/jails/jailname/root/usr/local/bin/stress