Should we add Lighttpd as an install/upgrade option?

Should we add Lighttpd as an install/upgrade option?

  • Yes, I would like to use lighttpd.

    Votes: 177 56.4%

  • I would like to use Nginx

    Votes: 66 21.0%

  • No, there are other things I would rather have.

    Votes: 71 22.6%
  • Total voters
    314
I WHOLEHEARTEDLY agree with Wido. I've been using DA since 8/03 ... exclusively. We've built our business for the last 5 years using it, but there are a TON of open issues that need to be addressed before going forward with lighthttpd

custombuild system needs a LOT of work, especially on upgrades
directory ownership issues
backup/restore issues
dns chroot issues
and the list of stoooopid little stuff goes on.

I beg you, don't make anther change / feature until these issues, and more, are cleared up.

Setting up new servers with the setup script on TOP of the basic DA install, (currently installing 40 boxes here) should NOT be this time consuming to make them have default features - roundcube, apf, els, security, webdav, CLAMAV, BFD, DDOS, dovecot, common CPAN scripts, spamassassin, php-imap, true-type, GD, ImageMagick, etc etc.

After DA base installs, we spend nearly an hour tweaking it to it's necessary levels before deploying as an active server. Let's fix what we've got, make it the best panel on the market (again) and THEN move forward.

Just my 2c.

Joe


Wido said:
I say no again.

I can't stress it enough, there are SO many bugs and issues to fix, please focus your energy on that first.

Choose a line to walk and get to the end of it, before adding new features.

I sum:

* Exim wich binds on port 80 with no reason
* Apache being restarted the whole time, wich causes some serious stability issues. DirectAdmin uses restart instead of graceful, wich kills all the active connections
* vm-pop3d and imapd who crash a lot and are seriously outdated. Dovecot should be more worked out and come as default, no more choices, use dovecot. The Maildir format is also much, much better then the Mbox format.
* The scripts directory is a total mess, no naming conventions at all, just a bunch of scripts dumped in there and hey, find it our yourself!
* The templates directory is the same, no explenation at all, and no naming convention at all.
* The API also has no conventions at all, features are added random without being tested fully

DirectAdmin should put much more time in perfectioning their current system, then expanding the features.

Documenting all the scripts en template files is one of them. Making naming conventions for all these files, dividing them in subdirectories.

Also, buffer the number of Apache restarts. Like, only restart Apache once every 10 minutes and use graceful instead of restart, that workes much better.

There are so many details to fix and i would love to sum them all when encountering them and submit them to the DirectAdmin team, but i have the feeling they don't have a real roadmap.

There also is no bugtracking system at all, so we don't know what the open issues are for example.

And before i am put away as a old grumpy man who has to much time, no, i am not.

I have to manage 100+ DirectAdmin servers and they consume SO much time, i sometimes regret i have chosing DirectAdmin, but on the other hand, i never found a good alternative.

No please, focus on the current issues first and then put your time in new features!
Click to expand...
 
I haven't been using DA as long as others but I will not only second this motion but tripple it and quadrouple it !!!! ...

Improve on fundamentals which are missing in the script right now before you start adding duplicate and triplicate features which most people don't need or want in the first place...

Improve the ability for Admins to turn ALL existing features on and off using the browser but most importantly, if an existing feature is turned off there is ABSOLUTELY NO NEED to tell the customer that the feature is available and it's turned off... Creates unnecessary support emails and explanations as to why specific features have been turned off...

Personally I thing you guy's should get your heads out of "geek mode" and get it into simple stupid "consumer mode" so you can eliminate the need for individuals to hunt down "How Do I Do? and WHere Do I Find?" support forums where the requests for "How Do I Do? and WHere Do I Find?" are repeated more than tenfold.... (Poor 'ol Jeff might stop sounding like a broken record for once...;))

This script is missing basic browser based necessities like the ability to move one user from one reseller account to another or the ability move ALL re-seller accounts form one to another (apparently being worked on now) and most importantly built in end user billing/accounting software which should have been incorporated into the script ages ago, eliminating the need for 3rd and 4th party sofware that is dependant on other servers or other parties to support...

I don't have to manage 40 servers but I am buying another server in the early part of the new year and I highly doubt that I will be buying another DA license unless I see some improvements on the basics that are already being offered in this script in the very near future...

I think I would rather pay a few bucks more to get something that is feature rich but a lot more efficient and simple stupid to use...

Another two cents worth...

Ed

PS: Sorry for getting off topic here , just venting.. :(


hostpc.com said:
I WHOLEHEARTEDLY agree with Wido. I've been using DA since 8/03 ... exclusively. We've built our business for the last 5 years using it, but there are a TON of open issues that need to be addressed before going forward with lighthttpd

custombuild system needs a LOT of work, especially on upgrades
directory ownership issues
backup/restore issues
dns chroot issues
and the list of stoooopid little stuff goes on.

I beg you, don't make anther change / feature until these issues, and more, are cleared up.

Setting up new servers with the setup script on TOP of the basic DA install, (currently installing 40 boxes here) should NOT be this time consuming to make them have default features - roundcube, apf, els, security, webdav, CLAMAV, BFD, DDOS, dovecot, common CPAN scripts, spamassassin, php-imap, true-type, GD, ImageMagick, etc etc.

After DA base installs, we spend nearly an hour tweaking it to it's necessary levels before deploying as an active server. Let's fix what we've got, make it the best panel on the market (again) and THEN move forward.

Just my 2c.

Joe
Click to expand...
 
Last edited:
Comparing the performance of things...

On my dual quad core clovertown, I ran some testing between the Nginx, Lighttpd, and Apache 2.2 worker mpm.


Static HTML (<html><head><title>Hello</title></head><body>Hello</body></html>):

Apache 2.2: 6650 requests/second 100% CPU Load
Lighttpd: 5670 requests/second 66% CPU Load
NGINX: 9320 requests/second 80% CPU Load

PHPInfo (<?php phpinfo(); ?>) - Running FastCGI on All

Apache 2.2: 465 requests/second 100% CPU Load
Lighttpd: 1820requests/second 100% CPU Load
NGINX: 2030requests/second 100% CPU Load


To me, it looks like Lighttpd is the most "light weight" but NGINX isn't much heavier and produces better performance for both PHP and Static HTML.


I think Apache might have performed better under mod_php rather than FastCGI, but I wanted to run the same php binary for all of the systems during the tests.

I also agree with the others, I don't think it's completely necessary at the moment, it might be better to perfect the other things as when you switch to NGINX you lose Perl support and I noticed even running the wrapper still caused issues. With lighttpd, perl support is there but you also lose things like rewrite and htaccess, etc.

Hope this helps...
 
I would have to say that I think that there are a lot of issues with DirectAdmin that could be addressed before addressing LigHTTPd, but I do think that it would be beneficial to your customers (e.g., me) who are actively seeking a solution to this.

If you add it, marvelous! If you don't, I hope you're working on improving other areas of DA. :)
 
I have been actually hunting for the best way to optimize web server and I just ran some tests today,

LiteSpeed totally destroys all other web servers (Apache/Litey/Nginx) mod_php/fastcgi, prefork or worker or not.

Only down side is that, its not free. It's like $800 per server.
They also have Free version but I don't know about that, i just tried their 15 day trial one.

LiteSpeed is really neat, they also have web-based gui to do all your php compiling etc....
 
if Litespeed cost the same as DA I would use it, but it's much more expensive. So I use NginX.
 
Agree with most... lighttpd would be something 'nice' to look at.. .but it shouldn't come at the expense of other things.

Also I think until all the functionality that many rely on with an apache webserver (.htaccess / mod_rewrite springs to mind) needs to be resolved cleanly first.

I would like to see in 'custombuild' that we have the option to enable/disable apache mods. (via options.conf) That is likely to drop the memory requirements of Apache considerably, and get much of the benefits that these alternate web-server solutions provides.

Just my $0.05 worth.
 
Personally I think it's going to be along time before prospects are going to sign up with a provider who advertises We use Litespeed vs a provider who advertises We use the latest version of the well-known Apache webserver.

But of course that's only my opinion :D.

Jeff
 
I think there is a problem here.

First the poeple voting are getting the same weight regardless of the amount of DA licences they use. I believe any small host with one server will more likely ask for a new fuzzy stuff like this because he would have only one or two server to update and is not facing the issues of DA that need improvement before new "fasthttpd" on a big scale.

Second, everyone using DA will agree that there are other most important features to add before loosing time with a new httpd daemon

Third, so many votes for a new server and so few replyies from the user that voted on this topic... it sounds odd to me.

So what is the position of the DA Staff ?
We haven't heard from them here since a long time...
 
Last edited:
I think that Lighthttpd should be an optional feature in DirectAdmin in conjunction with Nginx as well.
 
I disagree, I'm large lic holder but I too like to see httpd off from DA and see something better.



netswitch said:
I think there is a problem here.

First the poeple voting are getting the same weight regardless of the amount of DA licences they use. I believe any small host with one server will more likely ask for a new fuzzy stuff like this because he would have only one or two server to update and is not facing the issues of DA that need improvement before new "fasthttpd" on a big scale.

Second, everyone using DA will agree that there are other most important features to add before loosing time with a new httpd daemon

Third, so many votes for a new server and so few replyies from the user that voted on this topic... it sounds odd to me.

So what is the position of the DA Staff ?
We haven't heard from them here since a long time...
Click to expand...
 
I believe we all agree on the fact that another httpd client would be "nice" but don't you also think that there are other important features to fix/add before a new http daemon ?
 
netswitch said:
I believe we all agree on the fact that another httpd client would be "nice" but don't you also think that there are other important features to fix/add before a new http daemon ?
Click to expand...


my .02 cents I love DA, I won't buy anything else. I think they should increase the rate a little bit more and hire 1 more staff member working on the bugs and the other staff members working on new items. If money is the problem

It have been a slow process getting new items on board. Like i said, i love DA and I'm willing to wait, I'm hoping that 2008 things can speed up a bit tho

1. anything but apache
2. dual server setups (auto by DA)
3. client transfer from one DA to another
 
Well, DA needs a massive overhaul and try to make it more modern (with features of course)...ajaxy, web2.0 ish interfaces... you know very well, it's essential to stay in competition and have to evolve with the times.
 
agnivo007 said:
Well, DA needs a massive overhaul and try to make it more modern (with features of course)...ajaxy, web2.0 ish interfaces... you know very well, it's essential to stay in competition and have to evolve with the times.
Click to expand...

All that is just eye candy, DA can still have a basic look
 
I agree with all that say DA needs to be bug fixed first, before adding new applications.

Also a bug tracking system on this site would be a good feature! (Step 1 if you'd ask me!)

For those with a lot of servers, I'd like to request one more feature than already discussed. The update of DA itself, it must be a pain to have 100 DA servers and update them all. Wouldn't it be possible to have a "central update server", or a "master update server", that all "slave servers" could sync up with the master(s)?
 
agnivo007 said:
ajaxy, web2.0 ish interfaces...
Click to expand...

please NO ajax and other gimmicks in my control panel!

It should work with all the browsers out there and especially outdated ones. It's the system that controls the bare server, not a fancy social network gadget...
 
You must log in or register to reply here.
Back
Top