Use a rewrite rule so old bookmarks work regardless of domain
Zoot
Posts: 2,227
Jim Ewald, et al -- at risk of seeming presumptuous, my day job is as a web professional designing and deploying sites at a scale somewhat larger than parallax.com and forums.parallax.com, so I couldn't help mulling over some of the issues with the database conversion and wondering what *I* would on one of my own projects, given that I would not want to revisit the launched project and do it all over again :-)
There seem to be two *major* issues with the upgrade: munged old posts and the loss of usability in bookmarks (whether in a person's browser or in an index like Google), embedded links, etc.
It seems to me that a simple mod rewrite on forums.parallax.com will handle the second issue. Given that the old site (now at forums.parallaxinc.com) are all Active Server Pages (aspx files) and the new site (thank goodness) is PHP, all the mod rewrite rule would need to do is redirect any calls to any *.aspx* url to parallaxinc.com (i.e. just switching the domain name and preserving the rest of the query string). This would let old bookmarks, search indices, etc. work, and you would not need to move anything, rename any domains, translate old links, etc. In fact, even all the "cross-referenced" links embedded within thousands of old posts (whether viewed in the "new" forums or the "old") would work as well, w/o any need to retranslate the munged posts.
The only hitch I see is Jim's posted remark that the "old" forums would come down at the end of 2010. Is there a deep need to bring those pages down? Given that the old forums will not continue to grow, it would seem that leaving those page up indefinitely should a surmountable problem.
The first issue (munged posts) would seem to require a more substantial look at running those records through a php script to repair them; unfortunately, quick solutions there are probably not as achievable :-(
There seem to be two *major* issues with the upgrade: munged old posts and the loss of usability in bookmarks (whether in a person's browser or in an index like Google), embedded links, etc.
It seems to me that a simple mod rewrite on forums.parallax.com will handle the second issue. Given that the old site (now at forums.parallaxinc.com) are all Active Server Pages (aspx files) and the new site (thank goodness) is PHP, all the mod rewrite rule would need to do is redirect any calls to any *.aspx* url to parallaxinc.com (i.e. just switching the domain name and preserving the rest of the query string). This would let old bookmarks, search indices, etc. work, and you would not need to move anything, rename any domains, translate old links, etc. In fact, even all the "cross-referenced" links embedded within thousands of old posts (whether viewed in the "new" forums or the "old") would work as well, w/o any need to retranslate the munged posts.
The only hitch I see is Jim's posted remark that the "old" forums would come down at the end of 2010. Is there a deep need to bring those pages down? Given that the old forums will not continue to grow, it would seem that leaving those page up indefinitely should a surmountable problem.
The first issue (munged posts) would seem to require a more substantial look at running those records through a php script to repair them; unfortunately, quick solutions there are probably not as achievable :-(
Comments
But, I don't have your skillset. How hard would it be to knock something up?
This should be done as a matter of urgency to preserve existing content. Then once they sort out the formatting of old posts on the new vBulletin, they can just change the redirection.
Probably something like this ---
RewriteEngine on RewriteCond %{REQUEST_URI} !-U RewriteRule ^(.+) http://forums.parallaxinc.com/$1Actually, it could also be done with a .php error page (i.e. whatever page is called during a 404 error when a non-extant aspx file is requested). I would guess all of those links will return 404 errors in the server environment. In mentioning mod rewrite, I'm presuming that the PHP engine is running on an Apache-like *nix server, which may not be the case, as PHP is available for NT servers, and the use of ASP as a scripting engine on the old site is a clue that Parallax may favor Windows based servers, for whatever reasons.
E.g. something like:
<?php // this is a 404 error document on the "new" forums server... $str = "$_SERVER[REQUEST_URI]"; if ( preg_match("/forums\.parallax\.com/i", $str ) { header( "Location: ".preg_replace("/forums\.parallax\.com/i","forums.parallaxinc.com",$str) ); } else { header( "Location: http://forums.parallax.com/SomeExistingErrorPageFromForums.php" ); } ?>Or simpler, if everything is in the root of the old site:
I would always prefer to use mod rewrite, though. It's what that feature was designed for on modern *nix servers.
Thanks for taking the time to write up such a detailed solution on the redirects. The old forum site could remain online past the end of 2010 if needed. I'll have to move the database and IIS to different hardware but that is something that we can handle easily.
Your deductions are correct in that we are a M$FT shop making the transition to a Linux environment where we can. The Prallax Object Exchange was our first pure *nix deployment. This forum site was our second *nix project. Everything on this site is running under Linux.
We have a full development & test environment that mirrors this deployment. I can work on the rewrte rule in the morning.
The post formatting issue has been very elusive. I have tried numerous times to correct this within the vBulletin migration tool. The major difficulty is that the old site allowed posts to contain mismatched tags, tags that are out of order or missing end tags. The migration tool tried to make sense of it and in some cases made things even worse. I think the way forward on this issue is to identify specific errors and write a script to correct them.
Again, thank you for your suggestions. I'll report back tomorrow with the results.
Just in the post above there are two such code boxes for example. They look as ugly as hell on the page and to actually read the content, all of a copule of lines, one has to mess around with the horizontal and verticle scroll bars.
Does anybody think about us readers when implementing these useless features?
Consider Virtualizing the old forum. If it's held together with glue and duct tape, the most scary thing about that is preserving the state of it.
Once virtualized, this is a non issue. That old forum is very, very important, until this one gets sorted.
I don't think the importance can be over stated.
Virtualization, with tools such as VMWare, will result in the entire mess being a few files on some disk. Back that up, and the big mess is over. One can use a snapshot to test config changes and or scripts, for example. If it goes badly, just undo the snap shot, or pull the entire file from backup.
Take one home, if you want
"Consider Virtualizing the old forum."
Good idea, and whilst they are at it they can keep it active also. Then we can ignore all this junk and continue business as usual.
"That old forum is very, very important....I don't think the importance can be over stated."
Agreed, I don't know how someone could trash it so light heartedly.
"Take one home, if you want "
Yes please, how?
Personally, I would have done that a coupla years ago. The most scary thing about glue and duct tape is the current hardware state and how much of a PITA it is to recover it when bad things happen. This is understandable.
Virtualization is the answer to that, and I've been doing it to old, crotchety systems for years. It works, period.
Take one home = ftp big old zip file.
As for keeping it active. Well, I agree there too, and would have easily done it. Truth is, that's done, IMHO. (If I were them, I wouldn't undo where we are, but I would archive the old, and get to work making the new comfy.)
Also IMHO, the new software can be just fine. It's quick, and it's got a lot of options. Just need to work through those.
Most of the things mentioned are not a huge deal to me. Just change.
Code in scroll boxes is totally not ok. That's what the browser window is for! (clearly this has been forgotten in recent years)
And the default background, foreground color scheme isn't very readable. Most of the time, one is working in a moderate contrast environment. It would be good to either theme this and offer some choices, or go with very simple white, grey, black, and make darn sure the reading happens in a color space that makes sense. Often that's not as attractive, but then again... we are readers here, aren't we?
If the discussion is politics, cars, toys, or some other thing, perhaps it makes sense. Given the technical nature of the discussion here, some serious attention to the color scheme is warranted.
The only real issue with the virtual files going home is licensing... Always a mess, which is why OSS is such a good thing. A linux forum could be taken home.
Most of all, I need to hear the old forum will be maintained until this one is in a solid state to go forward. It's going to be hard to deal, if it's not, and that's just reality.
Just got done playing with some of the options. So, we get fonts!
And other fonts too!! Well, perhaps that's not a bad thing after all Maybe I'll just start posting in Lucida for a while..
Turning off the style support gets a person this (attached)
Never mind. I find it very, VERY frustrating to not be able to just upload a screen capture. What have I got, a few hundred on the old site? Grrr....
At least make the image defaults 800x600, so most captures will work. I WON'T resize them, nor link them off site. Too much work, and or the image quality changes are not ok.
Fine I'm up for it. Where is it?
How come I can't just do a recursive wget on the old forum and end up with a hard drive full of HTML and images and attachments ready to browse locally?
Something is very wrong with the world if one has to virtualize the entire file system, operating system and all just to view what is essentially text and linked images etc.
How are historians in the future going to do all of that?
No, I know the answer. In the future there is no past. It's all been unlinked, deleted and generally broken.
Problem is, there is a database, and a crotchety one at that, between us and a lot of work.
Sometimes that works.
Re: Historians.
Yes! I have a lot of concern over that because a lot of really crappy software and operating systems have been used contain a lot of human output.
This is true for many governments too, and I hate to bring politics here, but I will to react to your comment. When the work of the people --the civic work, not business which operates differently, is contained in ugly, proprietary things that must be paid for and that limit choices, what are we doing exactly?
Let me put one concern to rest. The old site will be available for as long as we need - longer if we can get it to a virtualized environment and avoid the hardware issues entirely.
That is good to hear. I believe in you guys. This is tough right now. Honestly, I had no idea how potent the "machine" was. Know that's what drives the comments, and that I think it's safe to say everybody wants us to get there again.
Rock on man! The battle probably begins Monday. Stay sane, and work through it. I, for one, appreciate it.
Good call on the VMWare. That's powerful, robust software. Highly recommended. Do buy into a copy of workstation though. You will find that license very handy.
______http://forums.parallax.com/showthread.php?t=124607
It still relies on images and attachments that remain on the old site, however. But these are just files, separate from the dotNetBB engine proper and could be relocated anywhere.
-Phil
LOL!! You've got serious job security right now. No worries right?
Yeah, I ran our little company on a small training laptop for about a week recently. We had a ugly outage, and I just moved them all to a nice big USB drive, fired off a copy of workstation, and had it all up and rocking in about an hour. Funny thing, is it all ran great. Accounting, CRM, and some test PLM server stuff.
Jim -- yer right -- sorry, my rewrite rule will capture ALL calls and redirect. But you get the idea.
Phil -- a good rewrite rule (or error page for that matter) should redirect traffic for ALL files, not just .html/.aspx/.php docs or the like, so that should cover for what would otherwise be "missing" images and such.