Is Mongrel dead?

rubyonrails-talk
1

Is Mongrel dead? When I look at http://mongrel.rubyforge.org/wiki/News the latest news was from close to a year ago, when last version of Mongrel was released. A look at the tickets shows a rather sad picture with only two developers “evanweaver” and “luislavena” contributing to the bug fixes lately.

Please understand that I am not complaining – I just want to know the situation so that I can make intelligent decision on the long term viability of Mongrel as an application server. I realize that it is a community project and if more people (including myself) would put in more effort things would be different. However, as it stands now, I am not able to contribute due to the lack of knowledge and time. This is a normal case of an open source software and, as I said, I am not complaining. Still, I would like to understand the long term implications of using Mongrel.

I know there is Phusion aka mod_rails. However, as it stands right now, the company is not a truly “for profit” company and relies on donations and consulting in installing mod_rails. Amount of dollars coming in for installation consulting is probably questionable since installation is fairly simple. The folks behind mod_rails hit the nail on the head however, when they introduced “Passenger Enterprise License”. Folks responsible for strategic long term decisions regarding technology need to be sure that the company won’t be left high and dry when Mogrel drowns (just look at that picture on http://mongrel.rubyforge.org/wiki - poor Fluffy is drowining). Rails proved to be a viable framework, that survived time test (although I would certainly prefer to see less frequent version releases). However, I am puzzled by the lack of interest in offering a commercial app server for Rails. There is certainly room for such a thing. Folks like me would rather pay XYZ dollars for a license, get phone support, etc. as oppose to get free Mongrel and keep my fingers crossed that new Ruby patch does not break it to pieces like it did last year. I know that there is a number of companies that use Rails regardless. However, the number would be bigger if there were more commercial tools, especially servers. If you look into Java world you will see that there is a reason why Weblogic successfully competes with free JBoss, Tomcat, and Glassfish.

…but I digress… So is Mongrel dead? Is there a commercial app server for Rails?

TPM

P.S. am intentionally posting this in the Rails forum as oppose to Mongrel forum or Deployment form, because I would like to hear from a wider audience and also let the wider audience see this thread. I think it is very important to the Rails community as it grows and as Rails tries to be on the par with Java and .Net in large corporations. Some indication where Ruby / Rails are can be gleaned from here: http://www.tiobe.com/index.php/content/paperinfo/tpci/index.html

2

Is Mongrel dead? When I look at http://mongrel.rubyforge.org/wiki/News the latest news was from close to a year ago, when last version of Mongrel was released. A look at the tickets shows a rather sad picture with only two developers “evanweaver” and “luislavena” contributing to the bug fixes lately.

Perhaps it is just good enough for most. Just because it isn't changing much doesn't mean that it is dead.

Please understand that I am not complaining – I just want to know the situation so that I can make intelligent decision on the long term viability of Mongrel as an application server. I realize that it is a community project and if more people (including myself) would put in more effort things would be different. However, as it stands now, I am not able to contribute due to the lack of knowledge and time. This is a normal case of an open source software and, as I said, I am not complaining. Still, I would like to understand the long term implications of using Mongrel.

EngineYard is still using mongrel and I have a client paying for slices that are using mongrel so it isn't like there's no backing for it.

I know there is Phusion aka mod_rails. However, as it stands right now, the company is not a truly “for profit” company and relies on donations and consulting in installing mod_rails. Amount of dollars coming in for installation consulting is probably questionable since installation is fairly simple. The folks behind mod_rails hit the nail on the head however, when they introduced “Passenger Enterprise License”. Folks responsible for strategic long term decisions regarding technology need to be sure that the company won’t be left high and dry when Mogrel drowns (just look at that picture on http://mongrel.rubyforge.org/wiki - poor Fluffy is drowining). Rails proved to be a viable framework, that survived time test (although I would certainly prefer to see less frequent version releases). However, I am puzzled by the lack of interest in offering a commercial app server for Rails. There is certainly room for such a thing. Folks like me would rather pay XYZ dollars for a license, get phone support, etc. as oppose to get free Mongrel and keep my fingers crossed that new Ruby patch does not break it to pieces like it did last year. I know that there is a number of companies that use Rails regardless. However, the number would be bigger if there were more commercial tools, especially servers. If you look into Java world you will see that there is a reason why Weblogic successfully competes with free JBoss, Tomcat, and Glassfish.

…but I digress… So is Mongrel dead? Is there a commercial app server for Rails?

TPM

I'm sure that if some held up a briefcase full of cash and asked if anyone was willing to exchange for some mongrel patches, you'd get some takers. Just make sure the amount of cash is reasonable given the kind of patch you expect to receive. :wink:

P.S. am intentionally posting this in the Rails forum as oppose to Mongrel forum or Deployment form, because I would like to hear from a wider audience and also let the wider audience see this thread. I think it is very important to the Rails community as it grows and as Rails tries to be on the par with Java and .Net in large corporations. Some indication where Ruby / Rails are can be gleaned from here: Home - TIOBE --

Just because a company wants to pay for supported software doesn't mean they understand the value of that software (or of OSS in general).

-Rob

Rob Biedenharn http://agileconsultingllc.com Rob@AgileConsultingLLC.com

3

Have a look at thin

http://code.macournoyer.com/thin/

It is easier to install and it uses the Mongrel engine as the core. I think this is where people that used to use Mongrel moved too.

4

Rob Biedenharn wrote:

Perhaps it is just good enough for most. Just because it isn't changing much doesn't mean that it is dead.

To me it does. I bet performance can be improved, a nice UI to manage Mongrels and view logs can be added, etc. There is a reason Apache did not stop at 1.2. If it did, we won't be using it now. Things do get outdated and constant improvement is necessary.

Just because a company wants to pay for supported software doesn't mean they understand the value of that software (or of OSS in general).

Agreed. However, the fact of life is that companies consist of soldiers, leutenants, colonels, generals, and politicians :slight_smile: soldier might know from experience that Rails/Mongrels is fine and his leutenant might agree. However, when it gets to the general's level he won't know and won't care - supplies to his troops must be guaranteed by a big corp (IBM, MSoft, HP, Sun, ...). This way politician can say something that sounds good to his constituents (shareholders). EngineYard is a good company, but it's no HP/Oracle/Some_other_heavy_weight_corp. In order for the Rails to be used at the highest corp level, there need to be backing from another big corp. There is a reason folks use Java and .Net to develop "ENTERPRIZE" level soft. I see how Sun is trying to get into the game with NetBeans and Glassfish capable of running JRuby/Rails. However, the effort seems to be just in case, kind of like MSoft's Iron Ruby. What I would LOVE to see is for HP or Oracle to step in and offer competition to MS/.Net and Sun/Java. I mean really back it up with a whole stack - app server, training, consulting, etc. It won't kill the little guys (there are plenty of Java consulting inc.'s besides the big 3), but it will get Rails to corp level for real.

5

Rob Biedenharn wrote:

Perhaps it is just good enough for most. Just because it isn't changing much doesn't mean that it is dead.

To me it does. I bet performance can be improved, a nice UI to manage Mongrels and view logs can be added, etc. There is a reason Apache did not stop at 1.2. If it did, we won't be using it now. Things do get outdated and constant improvement is necessary.

Oh, you think apache's updated UI is better?

sudo apachectl graceful

That's something mongrel should strive to achieve? It's hard to ignore the user-friendly Apache log viewer. Pardon the sarcasm, but sometimes the simpler the solution, the better. Thin is (to me) a slicker mongrel that is getting some attention. If there is a problem with mongrel, it's that some might have negative feelings about Zed's rant and connect that with mongrel (Zed's creation).

Rack has made servers much easier to create and there are a number of them available right now. I agree that Thin is a good drop-in replacement for mongrel. My limited tests show it consuming a bit less memory and serving requests a bit faster.

Just because a company wants to pay for supported software doesn't mean they understand the value of that software (or of OSS in general).

Agreed. However, the fact of life is that companies consist of soldiers, leutenants, colonels, generals, and politicians :slight_smile: soldier might know from experience that Rails/Mongrels is fine and his leutenant might agree. However, when it gets to the general's level he won't know and won't care - supplies to his troops must be guaranteed by a big corp (IBM, MSoft, HP, Sun, ...). This way politician can say something that sounds good to his constituents (shareholders).

Shareholders don't care so long as the company's stock is trading higher.

"Generals" who will only buy from IBM, MSFT, etc., *as you describe them* are unlikely to accept the terms of dealing with Rails. That said, there are a number of large corporations with Web sites hosted on OSS. Apache remains the leading deployment platform despite a lack of any "official" support from the vendor. It's not IIS, but that's a GOOD THING. If your "general" won't accept a solution built on an open source platform, that is a good thing to know up front so you don't chase your tail. But it really has more to do with the general than with mongrel.

EngineYard is a good company, but it's no HP/Oracle/Some_other_heavy_weight_corp. In order for the Rails to be used at the highest corp level, there need to be backing from another big corp. There is a reason folks use Java and .Net to develop "ENTERPRIZE" level soft.

Yeah, playing not to lose.

I see how Sun is trying to get into the game with NetBeans and Glassfish capable of running JRuby/Rails. However, the effort seems to be just in case, kind of like MSoft's Iron Ruby. What I would LOVE to see is for HP or Oracle to step in and offer competition to MS/.Net and Sun/Java.

They have an honest intellectual curiosity about dynamic languages, Ruby, and Rails. They also don't want to help create a situation where their platform becomes less relevant.

I mean really back it up with a whole stack - app server, training, consulting, etc. It won't kill the little guys (there are plenty of Java consulting inc.'s besides the big 3), but it will get Rails to corp level for real.

Ahhhhh. So we're back to the RMM issue. If you don't have a license, don't practice software development. That's a religious issue and if you feel better with "certified" trained developers, then buy into the proprietary stack. Again, you are less likely to be looking at a Rails solution because there is no single owner of a closed technology, and thus no accredited certifying authority that has natural credibility.

The MSFT certification program had mixed results. Naturally, because it was a barrier to entry, some talented developers became MCSEs. But some equally smart developers didn't. They were lost to any potential employer who used certification as a screening tool. Worse, it was possible to train for the certification exam, making the that a less valuable assurance of proficiency, as the developers who adopted this strategy were not necessarily the best and brightest.

This doesn't have anything to do with mongrel either.

Unasked-for advice: Before you hire a developer read his or her code. Have them explain their decisions. Then trust them. If you can't sell this upstairs, you've given up one of the strengths of OSS: The fact that you can read the code all the way down to the kernel if need be. And that is the best assurance of quality I can think of.

/RANT

6

I don't see mongrel going anywhere any time soon. True, Passenger is the way to go for production but for development mode mongrel is the clear winner. Who wants to setup an Apache instance and edit config files for every new project in development?

- Brian Cardarella

7

Hi,

I don't see mongrel going anywhere any time soon. True, Passenger is the way to go for production but for development mode mongrel is the clear winner. Who wants to setup an Apache instance and edit config files for every new project in development?

If you're on Mac OS X, you might want to take a look at: http://github.com/alloy/passengerpane

~ j.

8

As you point out, there's JRuby, which I've heard some larger "enterprise" type stuff is running on. I've even talked to a few folks who are sneaking Rails apps into their otherwise Java-only infrastructure that way.

However, you'll also find that a lot of folks have come to Rails (from Java/.Net) specifically to get away from the enterprise crowd. To use an animal analogy, Rails doesn't want to compete with the big dinosaurs - it's a small mammal, operating according to an entirely different set of rules.

--Matt Jones