Speeding up FireFox and Mozilla

[mobrien_12]

After updating FireFox and Mozilla on my home computer, I noticed that they were really really pokey.  They seemed to take unreasonably long times to resolve domain names, even though I could resolve them very quickly from the command line.  I spent a long time with google, and this is apparently a very common problem but the advice was all wrong (your networking is screwed up check this config file blah blah blah).  I finally found the answer.  

By going to about:config and changing "network.dns.disableIPv6"  from its default value of "false" to "true," the browsers stop trying to use IPv6 DNS for everything.  This makes a lot of sense because (GRRR) most people don't use IPv6.  So Mozilla/FF by default trys like blazes to use IPv6 DNS for everything, wasting time if it can't then falls back to what works.  

Everything got a lot faster for me when I threw this switch.  I'd be interested to know if it helps the rest of you.

[Dark_Me]

Interesting, why would it do that? Hmm... I'll try this on my box when I get home and tell you how it goes.

[Aloone_Jonez]

http://www.opera.com

[cymon]

That would also have the unintended side effect of bloating up Mozilla and Firefox.

[mobrien_12]

Well, I would hardly call it bloat...

IPv6 is a real standard. It will  replace IPv4.  It has to, simply because IPv4 doesn't have enough IP addresses in it to supply the needs of the future.  But, the future ain't here for most of us.

Apparently many programmers decided that they want their software to be IPv6 ready so that (a) their software can be dropped in an IPv6 environment today without any issues and (b) their software won't become obsolete because it's tied to a doomed standard.  The strategy that many have adopted is to have their software try IPv6 first, have it fail, then fall back on IPv4.  Well this is fine and dandy for things like MPlayer, where you don't make a new connection all the time, but for a  web browser it sucks.  

As for Opera, now it has IPv6 support (before it didn't).  However, Opera has IPv6 disabled by default (faster on the DNS resolves than a default firefox for IPv4, probably not for this switch thrown) so the people on IPv6 who want to use Opera are on the other side of the fence:  some of them bitch because Opera doesn't work at all out of the box and they want it configured like Firefox is!

I guess what it boils down to is that IPv6 is a really really good thing to have, but I think the Mozilla people made a huge mistake by making  it go IPv6 first, then fail down to IPv4 slowing down everybody but the small minority who have actually started to move to the new standard.  Second, they made a mistake by hiding this setting in about:config and not documenting it well (lots of people have been frustrated by it for years and almost nobody seems to know what to do).

Something with this much of an impact on speed should be in the Preferences control panel.  

By the way, if you do a search on google for this and find a post on a mozilla bbs with this solution in it by "guest999," that's me.  I got to many forum accounts.

I hadn't really noticed the slowdown on my work computer before, because it has a really fast CPU.  However, switching off the IPv6 DNS on firefox there sped things up there too... to the lightning fast level.

[cymon]

Excellent reply, but I was replying to Aloone_Jonez.

[mobrien_12]

Excellent reply, but I was replying to Aloone_Jonez.

Ah, well thanks

[Jack2000]

i can not see any major dif
between network.dns.disableIPv6's true and false settings

[Pathos]

If you don't have ipv6 capable hardware I would expect it to fail instantly and therefore not have any affect.

Do you notice how there is a gap between when you enter a url and the page title is shown and another gap before the content is displayed?

The first gap is the time it takes to get a dns reply, retrieve the html and process the html. The second gap is the time it takes to download CSS and javascript files and process any initial java script before rendering the page.

I find a huge reduction in the second gap when javascript is disabled because often offsite files are downloaded/processed by the initial javascript which takes time.

The other day Arts Technica was only displaying half the page because the javascript was stalling.

[Aloone_Jonez]

That would also have the unintended side effect of bloating up Mozilla and Firefox.

How?

I was joking, but Opera is an alternative if you get pissed off with Mozilla.

[davidnix71]

I'm using an eMac and Panther. Both the OS and hardware are IPv6 capable, but Mozilla's disable IPv6 value is/was "True." The only time I have problems is if I use v92 modem drivers. My isp doesn't use v92, so it becomes very unstable if I try.

I'm still using an older version of Firefox (1.0.4) because it seems every time I want to get the 'latest and greatest' most of my extensions won't work and I refuse to give them up.

[mobrien_12]

If you don't have ipv6 capable hardware I would expect it to fail instantly and therefore not have any affect.

Not from what I've seen.  

There was a big difference between my home and work computer.  My home computer was pissing me off.  I hadn't noticed it on my work computer.  

Your post made me dig deeper.  My home computer:  Fedora Core 4, which has IPv6 enabled by default.

My work computer:  WinXP (yuk) which did not have IPv6 installed (is not installed by default).  The speed up was there, but the slowdown was far far less significant.  

Do you notice how there is a gap between when you enter a url and the page title is shown and another gap before the content is displayed?
The first gap is the time it takes to get a dns reply, retrieve the html and process the html. The second gap is the time it takes to download CSS and javascript files and process any initial java script before rendering the page.

I noticed three distinct gaps. The first was the worst:  where it was taking forever while the info bar at the bottom said it was trying to resolve the URL.  The second would be your retrieve and process the html.  The third would be the CSS and javascript.  The second and third are not bad at all, usually because I kill javascript that pisses me off with adblock.

[Orethrius]

I'm using an eMac and Panther. Both the OS and hardware are IPv6 capable, but Mozilla's disable IPv6 value is/was "True." The only time I have problems is if I use v92 modem drivers. My isp doesn't use v92, so it becomes very unstable if I try.

Decent choices - if you have IPv6 capable hardware, other than shortening the load times slightly on IPv4 equipment, there's really no good reason not to have it enabled.  As for the modem, it's really a bad idea to run 56k.v92 on anything that can't handle it - just like you wouldn't run a cable line to a DSL modem.

I'm still using an older version of Firefox (1.0.4) because it seems every time I want to get the 'latest and greatest' most of my extensions won't work and I refuse to give them up.

Seriously?  Most of the extensions out there have 1.5.x counterparts, and those that don't - well, let's just say that the XPIs are actually ZIP archives, and you need to change one line of XML within said archive to fix most of them, and let you go from there.  

[davidnix71]

I've unzipped the xpi files and edited the version number, but I couldn't figure out how to rezip the file so Firefox would accept it as an extension. I have WinZip in Virtual PC and 7-zip. Could someone explain how to rezip the folder into an xpi after editing?

[inane]

Sweet