Re: [Mozilla Enterprise] Slow FF-ESR

[Luca Manganelli]

Here is it: https://bugzilla.mozilla.org/show_bug.cgi?id=1436297

Il 01/03/2018 09:23, Sylvestre Ledru ha scritto:

Hello Luca

Would you have the bug number ?

thanks

S


Le 01/03/2018 à 09:00, Luca Manganelli a écrit :

Il 28/02/2018 17:51, tagtraeumer ha scritto:

Hello,
if on FF-56.x esr, try to delete safebrowsing content (all files) in
folder:

C:\Users\xy\AppData\Local\Mozilla\Firefox\Profiles\xxx.yyy\safebrowsing

It did solve lame browsing issues for us on several cmputers...

Good luck + best regards,
Chris


I solved it better: disabiling COMPLETELY the safebrowsing module of
Firefox. Solved all 100% of problems (and, yes, I have opened a bug
report...)
___
Enterprise mailing list
Enterprise@mozilla.org
https://mail.mozilla.org/listinfo/enterprise

To unsubscribe from this list, please visit
https://mail.mozilla.org/listinfo/enterprise or send an email to
enterprise-requ...@mozilla.org with a subject of "unsubscribe"


___
Enterprise mailing list
Enterprise@mozilla.org
https://mail.mozilla.org/listinfo/enterprise

To unsubscribe from this list, please visit https://mail.mozilla.org/listinfo/enterprise 
or send an email to enterprise-requ...@mozilla.org with a subject of 
"unsubscribe"

Re: [Mozilla Enterprise] Slow FF-ESR

[Luca Manganelli]

Il 28/02/2018 17:51, tagtraeumer ha scritto:

Hello,
if on FF-56.x esr, try to delete safebrowsing content (all files) in folder:

C:\Users\xy\AppData\Local\Mozilla\Firefox\Profiles\xxx.yyy\safebrowsing

It did solve lame browsing issues for us on several cmputers...

Good luck + best regards,
Chris


I solved it better: disabiling COMPLETELY the safebrowsing module of 
Firefox. Solved all 100% of problems (and, yes, I have opened a bug 
report...)

___
Enterprise mailing list
Enterprise@mozilla.org
https://mail.mozilla.org/listinfo/enterprise

To unsubscribe from this list, please visit https://mail.mozilla.org/listinfo/enterprise 
or send an email to enterprise-requ...@mozilla.org with a subject of 
"unsubscribe"

Re: [Mozilla Enterprise] Slow FF-ESR

[Stephen Koppes]

I switched over to the x64 on my office computer about a year ago, then did the 
same on my personal computer shortly after. My personal computer is somewhat 
more powerful than my office computer, so I didn't notice the performance 
issues quite as rapidly as I did at the office.

FWIW, I'm not utilizing the ESR code-base at the current time on either my home 
or office machines, but am strongly leaning toward that as the replacement 
browser in our resource-constrained labs. It will be interesting to test 
because those machines are still 32-bit O/S.

Those RAM numbers are based on 58.0.2 x64 with Adblock Plus, Forecastfox (fix 
version), and Kee. Yes, I know that ABP is a resource hog, but with my testing, 
removing it didn't make the resource/freezing problem go away. In fact, the 
extra multimedia and advertisements caused the opposite - FF ran worse with all 
the extra website advertisement goodness loaded.


I'm happy to entertain any further questions regarding my experiences.

Stephen




On 2/28/2018 11:06 AM, Peter Saint-Andre wrote:


Have you experimented with Quantum (FF 57+) on another machine perhaps?
The main point of Quantum was to improve performance, but it would be
impossible to backport that work to the current ESR codebase. (FWIW I'm
running 1 FF Nightly window with 300+ tabs open, using ~1048MB RAM.)

Peter







Confidentiality Notice: This electronic communication and any attachments 
hereto contain confidential information from Penn Commercial, Inc. This 
information is only for use by the intended recipient and use by any other 
party is not authorized. If you are not the intended recipient, or believe that 
you have received this electronic communication in error, you are hereby 
notified that any disclosure, copying, distribution or the taking of any action 
in reliance on the contents of this communication or its attachments is 
strictly prohibited. If you have received this electronic communication in 
error, please delete it from your system and notify us immediately by telephone 
at (724)222-5330 or by reply electronic mail to the person from whom this 
message was received. Thank you.
___
Enterprise mailing list
Enterprise@mozilla.org
https://mail.mozilla.org/listinfo/enterprise

To unsubscribe from this list, please visit 
https://mail.mozilla.org/listinfo/enterprise or send an email to 
enterprise-requ...@mozilla.org with a subject of "unsubscribe"

[Mozilla Enterprise] Slow FF-ESR

[tagtraeumer]

Sorry, my fault:
FF-52.6x esr...

Hello,
if on FF-52.6x esr, try to delete safebrowsing content (all files) in
folder:

C:\Users\xy\AppData\Local\Mozilla\Firefox\Profiles\xxx.yyy\safebrowsing 

It did solve lame browsing issues for us on several cmputers...

Good luck + best regards,
Chris

___
Enterprise mailing list
Enterprise@mozilla.org
https://mail.mozilla.org/listinfo/enterprise

To unsubscribe from this list, please visit 
https://mail.mozilla.org/listinfo/enterprise or send an email to 
enterprise-requ...@mozilla.org with a subject of "unsubscribe"

[Mozilla Enterprise] Slow FF-ESR

[tagtraeumer]

Hello,
if on FF-56.x esr, try to delete safebrowsing content (all files) in folder:

C:\Users\xy\AppData\Local\Mozilla\Firefox\Profiles\xxx.yyy\safebrowsing 

It did solve lame browsing issues for us on several cmputers...

Good luck + best regards,
Chris

___
Enterprise mailing list
Enterprise@mozilla.org
https://mail.mozilla.org/listinfo/enterprise

To unsubscribe from this list, please visit 
https://mail.mozilla.org/listinfo/enterprise or send an email to 
enterprise-requ...@mozilla.org with a subject of "unsubscribe"

Re: [Mozilla Enterprise] Slow FF-ESR

[Peter Saint-Andre]

On 2/28/18 7:35 AM, Stephen Koppes wrote:
> Hello all,
> 
> I must also chime in on what has certainly felt like a significant
> performance loss that has crept into the FF engine. I went through the
> prior troubleshooting steps as well, but found the biggest 'problem' was
> continued use of the 32-bit version. Most all performance problems went
> away when I started utilizing the 64-bit version, so that might be worth
> a test?
> 
> In hind-sight, the explosive advent of multimedia driven sites, combined
> with add-on resource utilization, and the limitation of 2GB RAM per a
> 32-bit application was apparently more than it could readily handle with
> some of the more recent (~2 years) engine changes. Firefox was
> constantly freezing while running garbage collection for seconds at a
> time; while unfortunate, this is necessary, lest it try to use more RAM
> than a 32-bit application is allowed, and gets forcibly terminated by
> the O/S task scheduler.
> 
> 
> I'm concerned, however, that this is a band-aid style approach. With
> Firefox now running multiple processes with multiple tabs, and with the
> RAM usage increasing on some new builds, I'm concerned that Firefox may
> shortly be as bad as Chrome with resource utilization.
> 
> Right now I have 5 Firefox windows open, with 7, 8, 8, 4, and 3 tabs
> open, respectively. There are 7 instances of firefox.exe running, with a
> memory Private Working Set allocation ranging from 202MB to 698MB. Total
> 3,366MB private usage. I'm only actively using about 3 of these tabs;
> the others are for reference purposes, or projects for when I get more
> than a few spare minutes at a time.

Have you experimented with Quantum (FF 57+) on another machine perhaps?
The main point of Quantum was to improve performance, but it would be
impossible to backport that work to the current ESR codebase. (FWIW I'm
running 1 FF Nightly window with 300+ tabs open, using ~1048MB RAM.)

Peter




signature.asc
Description: OpenPGP digital signature
___
Enterprise mailing list
Enterprise@mozilla.org
https://mail.mozilla.org/listinfo/enterprise

To unsubscribe from this list, please visit 
https://mail.mozilla.org/listinfo/enterprise or send an email to 
enterprise-requ...@mozilla.org with a subject of "unsubscribe"

Re: [Mozilla Enterprise] Slow FF-ESR

[Stephen Koppes]

Hello all,

I must also chime in on what has certainly felt like a significant performance 
loss that has crept into the FF engine. I went through the prior 
troubleshooting steps as well, but found the biggest 'problem' was continued 
use of the 32-bit version. Most all performance problems went away when I 
started utilizing the 64-bit version, so that might be worth a test?

In hind-sight, the explosive advent of multimedia driven sites, combined with 
add-on resource utilization, and the limitation of 2GB RAM per a 32-bit 
application was apparently more than it could readily handle with some of the 
more recent (~2 years) engine changes. Firefox was constantly freezing while 
running garbage collection for seconds at a time; while unfortunate, this is 
necessary, lest it try to use more RAM than a 32-bit application is allowed, 
and gets forcibly terminated by the O/S task scheduler.


I'm concerned, however, that this is a band-aid style approach. With Firefox 
now running multiple processes with multiple tabs, and with the RAM usage 
increasing on some new builds, I'm concerned that Firefox may shortly be as bad 
as Chrome with resource utilization.

Right now I have 5 Firefox windows open, with 7, 8, 8, 4, and 3 tabs open, 
respectively. There are 7 instances of firefox.exe running, with a memory 
Private Working Set allocation ranging from 202MB to 698MB. Total 3,366MB 
private usage. I'm only actively using about 3 of these tabs; the others are 
for reference purposes, or projects for when I get more than a few spare 
minutes at a time.

Chrome is worse: My wife uses Chrome on her personal laptop and regularly hits 
the paging file with only 2 windows open and about a dozen tabs scattered 
between them. The laptop has 6GB of RAM, yet Chrome has at least 20 processes 
running, and in excess of 4.5GB total RAM utilization in the example above. 
Chrome obviously can't manage its resource utilization well enough to keep 
comfortably within that space. While Firefox is installed and up-to-date, she 
won't use it because she doesn't like the constant changes to work-flow.

I continue to use, support, and deploy Firefox at my office because we have a 
large quantity of resource-constrained machines. Chrome tests have always gone 
very poorly for the reasons stated above. I've stuck with older builds of 
Firefox because they work very well, given the technical constraints. Several 
sites are starting to not support the older version in our labs, I'm not 
looking forward to the next upgrade.


I'm not saying that Firefox shouldn't be innovating, but if we continue in a 
direction that appears we're copying Chrome, we're going to lose sense of 
uniqueness and BECOME Chrome. I don't feel that is good for anybody.







Stephen Koppes -- At Work
Network Administrator & Instructor
Penn Commercial Business/Technical School
242 Oak Spring Road
Washington PA 15301
Email: skop...@penncommercial.edu
Office: 724.222.5330 ext.338





Confidentiality Notice: This electronic communication and any attachments 
hereto contain confidential information from Penn Commercial, Inc. This 
information is only for use by the intended recipient and use by any other 
party is not authorized. If you are not the intended recipient, or believe that 
you have received this electronic communication in error, you are hereby 
notified that any disclosure, copying, distribution or the taking of any action 
in reliance on the contents of this communication or its attachments is 
strictly prohibited. If you have received this electronic communication in 
error, please delete it from your system and notify us immediately by telephone 
at (724)222-5330 or by reply electronic mail to the person from whom this 
message was received. Thank you.
___
Enterprise mailing list
Enterprise@mozilla.org
https://mail.mozilla.org/listinfo/enterprise

To unsubscribe from this list, please visit 
https://mail.mozilla.org/listinfo/enterprise or send an email to 
enterprise-requ...@mozilla.org with a subject of "unsubscribe"

Re: [Mozilla Enterprise] Slow FF-ESR

[Andrei Boros]

I tried clearing history, location history, download history. I haven't
went further because resetting profile also implies recreating the
entire work environment, not only bookmarks import, browser
configuration in detail, but also includes all existing addons, each
with it's own individual configuration, credentials (where applicable)
and customization and all site instances with saved passwords (not to
mention that passwords exporter addon cannot be compatible with Quantum,
if considering upgrades).

However I have encountered the same slugishness on new instances as
well, though not as severe in the beginning, definitely slowing down
with usage over time.

By comparison, SeaMonkey 2.46 and 2.49 with similar usage scenario,
though a times affected by the same problem, it is still much much
faster. The only problem are incompatible addons (FF only) and sites
that consider SM either as too old Firefox or incompatible browser.

Thank you.

On 2018-02-27 12:56 PM, William Spratt wrote:
>
> We experienced similar performance degradation, but it disappeared
> when reset profiles and set the new profile to autodetect network
> settings (not use the system settings – we’ve got an archaic proxy
> server that’s soon to be retired, and Windows 10 especially doesn’t
> seem to like talking to it; Firefox talks to it without issue).  We
> had been using a profile template (to set homepage, etc) but we’ve
> stopped using that. The only thing retained from the old profiles are
> bookmarks (by exporting to a bookmark backup and then importing into
> the new profile).
>
>  
>
> Have you tried doing something like that in your environment?
>
>  
>
> *Will Spratt*
>
> *IT Science Support Specialist*
>
> Tel: 01904 46 2631
>
>  
>
> *From:*Enterprise [mailto:enterprise-boun...@mozilla.org] *On Behalf
> Of *Andrei Boros
> *Sent:* 27 February 2018 10:39
> *To:* Sylvestre Ledru <sylves...@mozilla.com>; enterprise@mozilla.org
> *Subject:* Re: [Mozilla Enterprise] Slow FF-ESR
>
>  
>
>  
>
> On 2018-02-27 9:34 AM, Sylvestre Ledru wrote:
>
> Hello,
>
>  
>
> Le 27/02/2018 à 08:24, Andrei Boros a écrit :
>
> I have reported this bug:
>
> https://bugzilla.mozilla.org/show_bug.cgi?id=1424684 and the
>
> resolution was WONTFIX.
>
>  
>
> I have been witness to gradual end-user perceived performance decrease
>
> since FF-49 and it's gotten the worse at FF-52.
>
> Being at FF-52-ESR means being stuck with a barely usable browser and
>
> endless frustration.
>
> Downgrading is not much of an option because of profile format 
> changes.
>
> Upgrading to current stable, non-ESR branch is even less of an option,
>
> because of too many radical changes since FF57 and most importantly
>
> significant loss of functionality.
>
>  
>
> It also means now seriously considering alternativ forks such as
>
> SeaMonkey, PaleMoon, Waterfox and also other browsers altogether.
>
>  
>
> Why can't FF-52-ESR receive performance fixes, for performance
>
> degradation that appeared with it or it's recent predecessor versions?
>
> First, it is likely that Firefox 52 ESR is slower because of external
>
> factors (antivirus, security software, malware, etc) or addons.
>
> Given that I witnessed the overall slugishes on many installations in
> a variety of contexts, with and without external factors such as
> antivirus, security software, addons and lack of (new installations),
> my guess goes to FF-52-ESR, hence the bug report.
>
> In my tests I had to use real world profiles with real world addons.
> It is there where the performance impact is the problem. I could go so
> far as to clear history and also test with addons disabled, all but
> the profiling addon.
>
>
> We are extremely careful when modify Firefox ESR: we will only fix high
>
> and critical security issues, very important regressions due to
>
> new versions of operating systems or very important crashes.
>
> I can understand that and agree with you.
> However leaving it in a barely usable state ?
>
> The performance loss was gradual, since at least 49, and ongoing
> through 50 and 51, but it is in 52 (both release and esr) that it
> truly became a major hassle.
>
> We have data which shows that, improving performances in a stable
>
> product, leads in 50% of the case to regressions (stability, feature, 
> etc).
>
> Therefor, because we commit to stability to our users, we won't take any
>
> performance improvements into ESR.
>
>  

Re: [Mozilla Enterprise] Slow FF-ESR

[William Spratt]

We experienced similar performance degradation, but it disappeared when reset 
profiles and set the new profile to autodetect network settings (not use the 
system settings – we’ve got an archaic proxy server that’s soon to be retired, 
and Windows 10 especially doesn’t seem to like talking to it; Firefox talks to 
it without issue).  We had been using a profile template (to set homepage, etc) 
but we’ve stopped using that. The only thing retained from the old profiles are 
bookmarks (by exporting to a bookmark backup and then importing into the new 
profile).

Have you tried doing something like that in your environment?

Will Spratt
IT Science Support Specialist
Tel: 01904 46 2631

From: Enterprise [mailto:enterprise-boun...@mozilla.org] On Behalf Of Andrei 
Boros
Sent: 27 February 2018 10:39
To: Sylvestre Ledru <sylves...@mozilla.com>; enterprise@mozilla.org
Subject: Re: [Mozilla Enterprise] Slow FF-ESR


On 2018-02-27 9:34 AM, Sylvestre Ledru wrote:

Hello,



Le 27/02/2018 à 08:24, Andrei Boros a écrit :

I have reported this bug:

https://bugzilla.mozilla.org/show_bug.cgi?id=1424684 and the

resolution was WONTFIX.



I have been witness to gradual end-user perceived performance decrease

since FF-49 and it's gotten the worse at FF-52.

Being at FF-52-ESR means being stuck with a barely usable browser and

endless frustration.

Downgrading is not much of an option because of profile format changes.

Upgrading to current stable, non-ESR branch is even less of an option,

because of too many radical changes since FF57 and most importantly

significant loss of functionality.



It also means now seriously considering alternativ forks such as

SeaMonkey, PaleMoon, Waterfox and also other browsers altogether.



Why can't FF-52-ESR receive performance fixes, for performance

degradation that appeared with it or it's recent predecessor versions?

First, it is likely that Firefox 52 ESR is slower because of external

factors (antivirus, security software, malware, etc) or addons.
Given that I witnessed the overall slugishes on many installations in a variety 
of contexts, with and without external factors such as antivirus, security 
software, addons and lack of (new installations), my guess goes to FF-52-ESR, 
hence the bug report.

In my tests I had to use real world profiles with real world addons. It is 
there where the performance impact is the problem. I could go so far as to 
clear history and also test with addons disabled, all but the profiling addon.



We are extremely careful when modify Firefox ESR: we will only fix high

and critical security issues, very important regressions due to

new versions of operating systems or very important crashes.
I can understand that and agree with you.
However leaving it in a barely usable state ?

The performance loss was gradual, since at least 49, and ongoing through 50 and 
51, but it is in 52 (both release and esr) that it truly became a major hassle.


We have data which shows that, improving performances in a stable

product, leads in 50% of the case to regressions (stability, feature, etc).

Therefor, because we commit to stability to our users, we won't take any

performance improvements into ESR.



ESR60 will be released May 8th. Even if we recommend you wait until

ESR60.1 or .2, this version will benefit from all the Quantum improvements.
Well, it's not that long until then and normally I would contend with the wait, 
however the promised improvements come at a too great a cost in functionality 
loss. I have many users willing and wanting to stay at pre57 release and/or 52 
esr much longer with all risks involved.
Killing so much functionality with Quantum release was a move against FF users, 
whatever the stated reasons behind it.
I know many users who hava already ditched FF for Chrome or Opera because of 
it. Some of them still use FF release pre57 or esr52 for the specific 
functionalities and accompanying addons which gave FF a special edge compared 
to other browsers. But that is gone in Quantum, whatever new features it has.




Sylvestre



--

ing. Andrei Boros

Colectiv administrare-dezvoltare SITE
Radio Romania
Tel:   +40-21-303-1870
   +40-745-115721
Email: and...@srr.ro<mailto:and...@srr.ro>
___
Enterprise mailing list
Enterprise@mozilla.org
https://mail.mozilla.org/listinfo/enterprise

To unsubscribe from this list, please visit 
https://mail.mozilla.org/listinfo/enterprise or send an email to 
enterprise-requ...@mozilla.org with a subject of "unsubscribe"

Re: [Mozilla Enterprise] Slow FF-ESR

[Andrei Boros]

On 2018-02-27 9:34 AM, Sylvestre Ledru wrote:
> Hello,
>
> Le 27/02/2018 à 08:24, Andrei Boros a écrit :
>> I have reported this bug:
>> https://bugzilla.mozilla.org/show_bug.cgi?id=1424684 and the
>> resolution was WONTFIX.
>>
>> I have been witness to gradual end-user perceived performance decrease
>> since FF-49 and it's gotten the worse at FF-52.
>> Being at FF-52-ESR means being stuck with a barely usable browser and
>> endless frustration.
>> Downgrading is not much of an option because of profile format changes.
>> Upgrading to current stable, non-ESR branch is even less of an option,
>> because of too many radical changes since FF57 and most importantly
>> significant loss of functionality.
>>
>> It also means now seriously considering alternativ forks such as
>> SeaMonkey, PaleMoon, Waterfox and also other browsers altogether.
>>
>> Why can't FF-52-ESR receive performance fixes, for performance
>> degradation that appeared with it or it's recent predecessor versions?
> First, it is likely that Firefox 52 ESR is slower because of external
> factors (antivirus, security software, malware, etc) or addons.
Given that I witnessed the overall slugishes on many installations in a
variety of contexts, with and without external factors such as
antivirus, security software, addons and lack of (new installations), my
guess goes to FF-52-ESR, hence the bug report.

In my tests I had to use real world profiles with real world addons. It
is there where the performance impact is the problem. I could go so far
as to clear history and also test with addons disabled, all but the
profiling addon.

> We are extremely careful when modify Firefox ESR: we will only fix high
> and critical security issues, very important regressions due to
> new versions of operating systems or very important crashes.
I can understand that and agree with you.
However leaving it in a barely usable state ?

The performance loss was gradual, since at least 49, and ongoing through
50 and 51, but it is in 52 (both release and esr) that it truly became a
major hassle.
> We have data which shows that, improving performances in a stable
> product, leads in 50% of the case to regressions (stability, feature, etc).
> Therefor, because we commit to stability to our users, we won't take any
> performance improvements into ESR.
>
> ESR60 will be released May 8th. Even if we recommend you wait until
> ESR60.1 or .2, this version will benefit from all the Quantum improvements.
Well, it's not that long until then and normally I would contend with
the wait, however the promised improvements come at a too great a cost
in functionality loss. I have many users willing and wanting to stay at
pre57 release and/or 52 esr much longer with all risks involved.
Killing so much functionality with Quantum release was a move against FF
users, whatever the stated reasons behind it.
I know many users who hava already ditched FF for Chrome or Opera
because of it. Some of them still use FF release pre57 or esr52 for the
specific functionalities and accompanying addons which gave FF a special
edge compared to other browsers. But that is gone in Quantum, whatever
new features it has.
> Sylvestre
>

-- 

*ing. Andrei Boros*

Colectiv administrare-dezvoltare SITE
*Radio Romania*
|Tel:   +40-21-303-1870
   +40-745-115721
Email: and...@srr.ro |

___
Enterprise mailing list
Enterprise@mozilla.org
https://mail.mozilla.org/listinfo/enterprise

To unsubscribe from this list, please visit 
https://mail.mozilla.org/listinfo/enterprise or send an email to 
enterprise-requ...@mozilla.org with a subject of "unsubscribe"

[Mozilla Enterprise] Slow FF-ESR

[Andrei Boros]

I have reported this bug:
https://bugzilla.mozilla.org/show_bug.cgi?id=1424684 and the resolution
was WONTFIX.

I have been witness to gradual end-user perceived performance decrease
since FF-49 and it's gotten the worse at FF-52.
Being at FF-52-ESR means being stuck with a barely usable browser and
endless frustration.
Downgrading is not much of an option because of profile format changes.
Upgrading to current stable, non-ESR branch is even less of an option,
because of too many radical changes since FF57 and most importantly
significant loss of functionality.

It also means now seriously considering alternativ forks such as
SeaMonkey, PaleMoon, Waterfox and also other browsers altogether.

Why can't FF-52-ESR receive performance fixes, for performance
degradation that appeared with it or it's recent predecessor versions?
And by that I don't mean the large-scale redesign of 57 and later, but
rather adequate fixes for 52 generation.

Thank you
-- 

*ing. Andrei Boros*

Colectiv administrare-dezvoltare SITE
*Radio Romania*
|Tel:   +40-21-303-1870
   +40-745-115721
Email: and...@srr.ro |

___
Enterprise mailing list
Enterprise@mozilla.org
https://mail.mozilla.org/listinfo/enterprise

To unsubscribe from this list, please visit 
https://mail.mozilla.org/listinfo/enterprise or send an email to 
enterprise-requ...@mozilla.org with a subject of "unsubscribe"