Mailing Lists: Apple Mailing Lists

Image of Mac OS face in stamp
 
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: Trouble installing custom php 5 on 10.5.4 with mcrypt




On Jul 26, 2008, at 00:46 AM, Brian Blood wrote:


On Jul 25, 2008, at 9:52 PM, webmaster wrote:

I did this by using the lipo command to thin out Apple's apache2 binary
(after making a backup of course) to just contain a 32-bit x86 image.


Works like a charm.

This is not a recommended path to follow, you're better off compiling things using the correct flags than you to open yourself up to issues caused by software update.



I guess you've haven't grasped the universal constant when it comes to judging the efficacy of technological solutions:


It Depends.

If you're running client as a weekend warrior and not a deployed server in a production environment then I would say breaking it due to an OS update isn't mission critical.


The OS was supplied with 64bit capabilities for a reason and in a real server environment it makes practical sense to utilize the advantages offered from running in a 64bit environment.


YOUR situation may compel you to decide that this is not the a "recommended path", however there are others in the world who have differing circumstances where this particular approach works out absolutely brilliantly.

Off-hand I don't know of any real deployed production servers that use a thinned apache since the benefits of running in 64bit are much more attractive and this is the general consensus and not my personal situation.


Case in point, the server I first did this on was in need of being updated from 10.5.2 to 10.5.4.
I ran the update and yes, Apple HAD updated the httpd binary wiping out my thinned out binary. Apache was not loading due to the expected errors described previously.

If you did it right it wouldn't have broken so your "Case in point" isn't really valid, it's just a lazy way to get it up and running with software generated by someone else that wasn't built for or intended to be used in a real server environment, entropy is an example of this and I have discussed this with Marc many times in the past, he does it this way out of conveniences, not out of requirements.


No matter. I had placed my commands for thinning httpd down into a shell script. I simply reran that and lo and behold launchd pretty much immediately started up the new i386-only based httpd without me having to do anything else.

Problem solved. Case closed. No muss. No fuss.

The fact that it broke due to an OS update is reason enough to not do it on a deployed server in a production environment unless you have absolutely no other choices or options and the benefits would have to outweigh the cons considerably and that is why it's not a recommended procedure.


The better solution would have been to build your add-on software to use both architectures and you wouldn't have had any issues when software updated occurred now or later in the future so there was fuss and muss because the update broke your server.




Indeed. I guess I was just being 64bit stubborn ;)

Not really, right idea wrong approach.

Well, in my case, this particular servers usage pattern was such that I deemed this an acceptable workaround that is well documented in our system and works flawlessly. If Jeremy deems it works for him and he accepts the consequences, then good for him.

Suggesting such an approach should have been as a last resort when no other options are available, doing it the right way is what I recommend to avoid any future issues, if he wants to take a different approach that's up to him, I offered my recommendations and information to help achieve the required results which aren't susceptible to OS updating failures.


Looking at your other email I see you recommend spending orders of magnitude more time recompiling for an end result that most likely provides no real-world practical benefit.

That is not so, I recommend spending the time planning and doing it right the first time so you don't have to worry about rebuilding it or fixing it later cause it breaks due to an OS update.


So how is thinning out the binary and being back up in 5 seconds the "wrong" approach?

If you have to ask then it's obvious you don't understand why.


Dale, I respect your knowledge in how to tear apart and rebuild the modules in PHP. You've frequently shared valuable knowledge in that area with the list.


However in this case, a better approach might have been:

"Thinning out the binary is an interesting approach. It does pose deployment risks from a Software Update wiping out your change, but if you REALLY wanted 64-bit, here's how to do it:"

Knowing a little about his setup and his use is the reason I recommend he do it properly now rather than placing his server in a heightened position of potential failure due to an OS update.


Otherwise, you just come across as rude and arrogant.

If you interpret it as rude that's OK, others probably share the same opinion and everyone is entitled to their own.




Brian

-- Dale


_______________________________________________ Do not post admin requests to the list. They will be ignored. Macos-x-server mailing list (email@hidden) Help/Unsubscribe/Update your Subscription: http://lists.apple.com/mailman/options/macos-x-server/email@hidden

This email sent to email@hidden
References: 
 >Trouble installing custom php 5 on 10.5.4 with mcrypt (From: Jeremy Bush <email@hidden>)
 >Re: Trouble installing custom php 5 on 10.5.4 with mcrypt (From: Brian Blood <email@hidden>)
 >Re: Trouble installing custom php 5 on 10.5.4 with mcrypt (From: Jeremy Bush <email@hidden>)
 >Re: Trouble installing custom php 5 on 10.5.4 with mcrypt (From: webmaster <email@hidden>)
 >Re: Trouble installing custom php 5 on 10.5.4 with mcrypt (From: Brian Blood <email@hidden>)



Visit the Apple Store online or at retail locations.
1-800-MY-APPLE

Contact Apple | Terms of Use | Privacy Policy

Copyright © 2007 Apple Inc. All rights reserved.