I’m not sure if by design, however this was not listed as a change that I remember seeing.
Second issue:
After core 188 update CPU usage has risen slightly. This in itself is a minor increase but in the low-traffic scenario that I have - if such an increase was mirrored in a bigger environment that could be an issue:
The increased usage is almost exclusively additional ‘User’ CPU time usage and I’d say it’s roughly 2-3x higher than it was previously.
The change mentions changing QoS to decrease CPU usage - maybe the different software method has a slightly higher baseline usage with less additional overhead, or more likely some other software component is doing this…
Note: this additional usage is not from additional / heavy traffic - as seen from time shown in graph, the update was done very out of hours.
I did want to check the processes graph to see if a service was using more but that’s MIA since updates.
EDIT: After doing a bit of investigation it would appear that CPU usage of the httpd service when connected to the GUI via browser accounts for the extra CPU usage. This is strange as the browser session was open for some time prior to the update and the CPU usage remained flat.
Indeed, but curious how it’s measurably worse now somehow - it’s not like it’s doing a lot more compared to previous version
To be honest I was initially more concerned it may have been an issue with some core process which may have been amplified for high traffic environments but fortunately not.
That said the missing services process graph doesn’t help - whilst it was little issue for me to log in to console as root and run top, there may be some users who aren’t familiar enough with linux to know that (but then they probably wouldn’t have cared).
Those screenshots were taken from menu System->System. Is this what you were after?
No. That just shows the load average across CPU(s). The process graphs from the services page shows the individual process resource usage over time for things like OpenVPN, squid, guardian/IPS, etc.
It’s a pretty simple question to answer from the devs; “yes it should be there but isn’t” or “oh, we took that out and didn’t mention it”.
You are correct. I logged into a CU187 system and if you scroll down below Services and Addon - Services, there are two graphs: Processes Graph and Processes Memory Graph. Those are both missing from my CU188 system.
Also, IMO the formatting on CU188 looks less professional. The Status column is removed and integrated into the two remaining columns with a large red bar “STOPPED” that throws the symmetry of the page off.
Also, IMO the formatting on CU188 looks less professional. The Status column is removed and integrated into the two remaining columns with a large red bar “STOPPED” that throws the symmetry of the page off.
Yes, those two graphs were removed in the CU188 update. Those parameters were not a single measurement but a combination of several factors and basically, when it went high or when it went low it was very difficult to determine what it was actually indicating was happening with the IPFire system.
The Processes graph also did not include all processes running on IPFire.
I had no idea how to interpret a change during a day from running 1,000 processes a day to running 30,000 processes a day. Squid always seemed to dominate it anyway as it has a large number of child processes that get started so it will typically swamp the others that are included.
The Processes Memory Graph I have never understood, and always look at the Memory graph if I have concerns about it.
I do not believe it provided any useable information to take action on.
In future, if something is removed from a wui page we will try and remember to explicitly mention it.
