Posted by "Richard Lynch" on 08/24/05 09:42

On Tue, August 23, 2005 9:44 am, Rick Emery wrote:
> Quoting Rick Emery <rick@emery.homelinux.net>:
>
>> Ugh, we're *never* going to make a decision. My boss just sent me
>> this email:
>
> I wrote an application, using PHP5, that displays a list and refreshes
> every 30 seconds (the data is constantly changing, but a 30 second
> delay is acceptable). As I've indicated previously, we're a Microsoft
> shop, so the data comes from MS SQL Server 2000. No problems, the app
> worked great using my workstation as the server with a few clients
> running the app from it. It even worked when we moved it to a server
> and opened it up to everyone on our intranet (for a while).
>
> We have two different types of clients. Some use desktop computers,
> physically connected to our network, while others use mobile laptops
> connected to our network via cellular (using Sprint AirCards) using
> third-party VPN software (Padcom, in case anyone's familiar).
>
> We set the application up on a Windows 2000 Server with IIS (5, I
> think), and it would work fine for about a day. Then Padcom clients
> kept stopping. They'd request the page and, after a loooooong time,
> display a message that the request timed out. This would seemingly
> happen for all Padcom-connected clients at the same time, though the
> desktops continued to work fine. We restarted the server running the
> Padcom software with no effect. We restarted IIS on the web server
> with
> no effect. The only thing (seemingly) that cleared the issue was
> rebooting the server running IIS.

Just for a test case, write a 10-line ASP script that does something
similar, if much simpler, and pound on it on the same box with the
Padcom clients.

I'm betting you'll have the SAME ISSUE, and that the problem has
NOTHING to do with PHP whatsoever.

PHP works fine with IIS and Windows.

Or, rather, PHP doesn't make IIS and Windows any LESS stable than they
already were without PHP.

> has mentioned changing anything, until this morning. My manager
> informed me in our meeting that no language could be chosen unless it
> works under IIS.

I'll say it again: There is *NOTHING* wrong with PHP and IIS.

IIS and Windows are badly-broken, all on their own, without PHP.

> So, I'm faced with finding an obscur problem, running on obscur
> software (the vendor for Padcom, of course, insists that they've never
> seen this problem). I'm confident that the problem has *nothing* to do
> with PHP, but am forced by management to try to prove it.

If you can prove it, showing the same problem in ASP, then what?

Are they going to let you move the application back to a real server?

In fact, you could probably get ahold of a Padcom and prove it to
yourself in a days' work, and then get them to agree that if it's not
PHP nor your script that's broken, but Windows+IIS, then maybe they
should just leave the WORKING stuff alone.

Probably won't work.

But that's how office politics work.

--
Like Music?
http://l-i-e.com/artists.htm

[Back to original message]