Re: [Evolution] Help me understand something please

[Pete Biggs <pete biggs org uk>]

On Sun, 2022-06-26 at 12:37 -0500, tmcconnell168 gmail com wrote:
> On Sat, 2022-06-25 at 23:21 +0100, Pete Biggs wrote:
> > > I was clicking group reply and not using the drop down next to it,
> > > Apparently it's not the same thing. 
> >
> > Edit -> Preferences -> Composer Preferences ->
> >        Group reply goes only to mailing list if possible.

     ^^^^^^
     This, Please.

> >
> Only one small problem with Logwatch, it sends a mail like what is
> having the high CPU usage when I try to view them. It's those types of
> mails that are causing high CPU usage. 

I get logwatch emails from about 100 machines. Some of them are over
250k and I have never had any issues - there is a delay while the mails
are downloaded over IMAP, but the rendering does not result in high CPU
usage. 

I get kernel MCE events from them as well.  Some of them are nearly 1Mb
and again it doesn't seem to be an issue.

> Those types of reports, from
> Cron daily, Cron Weekly, Tripwire -m c (Cron Daily), Debian Security
> Status (I would have NO control over the number of Vulnerabilities
> found/solved) Unattended Upgrades (I have no control over the number of
> bug fixes released) which tells success/failure and what was upgraded. 

I have one debian system sending me those sort of reports. Again no
issues with any of the emails.  

What's also clear is that no one else is seeing any issues like this.
Which points to either a configuration/software issue or a hardware
problem.

Configuration issues can be eliminated by creating a separate Linux
user and using that to read the emails to see if it causes problems.

Software can be dealt with by temporarily swapping out the disk(s) and
re-installing the OS on new disks.

Similarly hardware can be investigated by using different hardware.

I understand that not everyone has the resources to do this, but to get
to the bottom of this you need to eliminate some of the variables.


> Okay here's the 5 backtraces from WebKitWebProcess. 
If you haven't already, then create a Gnome/Evolution bug report and
attach the backtraces to there - it's more likely that it will get
traction from the people who really know what they're doing!  Could you
put the Evolution bug report numbers in this thread so that anyone who
does have similar problems in the future can find them.


> 6. run `sudo kill PID` for each WKWP listed and restart Evolution (and
> possibly machine, success varies with this). 

Why sudo?  The processes should be owned by your username.

P.