On 30 Oct 2025 at 6:57, Roger Heflin wrote:

From:                   Roger Heflin <[email protected]>
Date sent:              Thu, 30 Oct 2025 06:57:11 -0500
Subject:                Re: Strange 30 second delay starting chromium or 
chrome with regular
        user, but not with root??
To:                     [email protected],
        Community support for Fedora users 
<[email protected]>
Send reply to:          Community support for Fedora users 
<[email protected]>

> On Wed, Oct 29, 2025 at 4:55 PM Michael D. Setzer II via users
> <[email protected]> wrote:
> >
> > Recently seen a delay of around 30 seconds with both chrome and
> > chromium on Fedora 42? Before both would be almost instant.
> >
> > line
> > ppoll([{fd=56, events=POLLIN}], 1, {tv_sec=24,
> > tv_nsec=999962000}, NULL, 8) = 0 (Timeout)
> >
> > Seems fd=56 is linked to libscret?
> >
> > openat(AT_FDCWD, "/lib64/libsecret-1.so.0",
> > O_RDONLY|O_CLOEXEC) = 56
> 
> A Ppoll call makes no sense to be used/timeout for a file/library.

I've used strace with command about 12 times, and always seem 
the same things.
From the latest time. 
Running strace chromium-browser 2>/tmp/outz
and closing it once it opens  created a file with 35657 lines.
Find lines
10036   openat(AT_FDCWD, "/lib64/libsecret-1.so.0", 
O_RDONLY|O_CLOEXEC) = 56
10057   ppoll([{fd=56, events=POLLIN}], 1, {tv_sec=24, 
tv_nsec=999996000}, NULL, 8) = 0 (Timeout)

So it always seems to be the 56 is linked.

Running script with regular use just got
Should take about 10 seconds to 1 minute to download and process
Took 55 Seconds to Download

Should take about 10 seconds to 1 minute to download and process
Took 7 Seconds to Download

Same exact code.
browser=chromium-browser
cp emplistc2.csv emplistc2.bak
clear
echo "Should take about 10 seconds to 1 minute to download and 
process"
rm ./Campus*.pdf ./Campus*.txt -f
base="https://employeessb-prod.ec.guamcc.edu/EmployeeSelfServi
ce/ssb/campusDirectory#/lastName"
opt="--no-sandbox --virtual-time-budget=333 --disable-pdf-tagging 
--headless --print-to-pdf=Campus"
touch Campus1.pdf
start_time=$SECONDS
$browser $opt"1.pdf" "$base"                                      2>/dev/null &
$browser $opt"2.pdf" 
"$base?nextRow=56&previousRow=1&page=2"      2>/dev/null &
$browser $opt"3.pdf" 
"$base?nextRow=111&previousRow=56&page=3"    2>/dev/null &
$browser $opt"4.pdf" 
"$base?nextRow=166&previousRow=111&page=4"   2>/dev/null &
$browser $opt"5.pdf" 
"$base?nextRow=221&previousRow=166&page=5"   2>/dev/null &
$browser $opt"6.pdf" 
"$base?nextRow=276&previousRow=221&page=6"   2>/dev/null &
wait
echo "Took "$((SECONDS - start_time))" Seconds to Download"

Use to download pages with wget2 in about 6 seconds, but vendor 
changed page info and that no longer works. Using firefox can 
manually go to each page and use print-to-pdf to get real data for 
each page. Using save as gives a strange 1.3M java script result 
instead of screen data? firefox does have option to pdf via 
command line.  That takes about 2 minutes.

base="https://employeessb-prod.ec.guamcc.edu/EmployeeSelfServi
ce/ssb/campusDirectory#/lastName"
firefox $base
mv Campus\ Directory.pdf Campus1.pdf
firefox "$base""?nextRow=56&previousRow=1&page=2"
mv Campus\ Directory.pdf Campus2.pdf
firefox "$base""?nextRow=111&previousRow=56&page=3"
mv Campus\ Directory.pdf Campus3.pdf
firefox "$base""?nextRow=166&previousRow=111&page=4"
mv Campus\ Directory.pdf Campus4.pdf
firefox "$base""?nextRow=221&previousRow=166&page=5"
mv Campus\ Directory.pdf Campus5.pdf
firefox "$base""?nextRow=276&previousRow=221&page=6"

Opens each page, and do Ctrl-P pause Enter pause Enter pause 
Ctrl-Q.

All process download the 6 web pages, print to PDF, then use 
pdftotext to get about 50K text file each. Extract data takes
Total Time to Down/Process/Diff 0.006698 Seconds

Gets about 248 records of data from Campus Directory.

Just interesting. 
Thanks again.

> 
> You have to be careful with fds, they get closed and reused all of the
> time, so likely there was an earlier open of 56 and a close of 56
> before the openat you have.
> 
> So you need to find what call opened 56 before the ppoll timeout.   It
> is probably going to be some network socket based on ppoll being used.
> 
> And given google chrome does DUMB things (like complaining that it
> cannot UPDATE its code-big shocker since it is installed as ROOT but
> running as a USER and was installed from an RPM).    Whoever coded it
> for Linux ASSUMES that Linux works like Windows.
> -- 
> _______________________________________________
> users mailing list -- [email protected]
> To unsubscribe send an email to [email protected]
> Fedora Code of Conduct: 
> https://docs.fedoraproject.org/en-US/project/code-of-conduct/
> List Guidelines: https://fedoraproject.org/wiki/Mailing_list_guidelines
> List Archives: 
> https://lists.fedoraproject.org/archives/list/[email protected]
> Do not reply to spam, report it: 
> https://pagure.io/fedora-infrastructure/new_issue


+------------------------------------------------------------+
 Michael D. Setzer II - Computer Science Instructor (Retired)     
 mailto:[email protected]                            
 mailto:[email protected]
 mailto:[email protected]
 Guam - Where America's Day Begins                        
 G4L Disk Imaging Project maintainer 
 http://sourceforge.net/projects/g4l/
+------------------------------------------------------------+



-- 
_______________________________________________
users mailing list -- [email protected]
To unsubscribe send an email to [email protected]
Fedora Code of Conduct: 
https://docs.fedoraproject.org/en-US/project/code-of-conduct/
List Guidelines: https://fedoraproject.org/wiki/Mailing_list_guidelines
List Archives: 
https://lists.fedoraproject.org/archives/list/[email protected]
Do not reply to spam, report it: 
https://pagure.io/fedora-infrastructure/new_issue

Reply via email to