dokether

Awesome. Thank you!

Trying to fix the feed in the meantime so I have delete the tweet that is causing the issue. You can see a copy of the tweet here:

https://cl.ly/image/2y3Q23310o28

Ipstenu,

Sorry, I forgot to respond to close this up. After hammering the tech support at my host for a while they clued in that they had a 301 Redirect setup for all domains:

The 301 redirect was coming from our configuration,
as for all sites on Flywheel (except subdomain
multisite) we redirect to the primary domain. This
is the correct behavior, even for subdirectory based
mulitsite, but in this case you’re also mapping
domains so that’s not ideal.

That is correct…

Site ID: 2
Domain: theowner.com
Primary: yes

Right now when I type in “theowner.com”, I get redirected to domain.com. It’s like WordPress doesn’t realize the MU Domain Mapping plugin is installed which seems really weird to me. I also have no redirection plugins or anything turned on at the moment.

Thank you!

So that’s what it was, the mail server was configured wrong and was causing issues. In case someone needs in the future, this was the final message from support detailing what was wrong:

—
The issue was the server hostname was kether-interactive.ca and you were sending an email to [email protected] so it thought it was local but couldnt send to the local system since mail wasnt running on it.
—

They have the e-mail for kether-interactive.ca setup on a different server than my dedicated hosting server, so it was throwing an error everytime because it was trying to send to the local mail system. A rather obscure error but the answer is here in case anyone else runs into a similar issue.

I think they have the path for the php sendmail function wrong. I realized that creating sites, deleting/creating users, and even MarketPress test purchases weren’t sending me the e-mails that they should have been.

Setup a basic php mailer:

<?php mail("[email protected]", "Testing E-Mail", "Testing E-Mail Contents", "From: [email protected]"); ?>

500 Internal Server Error. ??

Heh, commented out the line and still getting the 500 Error so hopefully they’ll finally decide to look deeply into their server configuration haha.

Their response:
—
Due to the server configuration there is no way to change it to the apache php module. If we didnt you wont have any page load, it would be 500 errors no matter what page you went to. Also I think its more along the lines of a wordpress issue, since cPanel uses php in CGI mode for security, and there are more cPanel servers running the wordpress installs running, which you can see here:

https://supportcenter.verio.com/KB/questions.php?questionid=316

“Default PHP mode: Since SuPHP is used by default, PHP runs in CGI mode. “

This is why I asked to remove the code in the file since this then takes out the equation that its the server. If taking it out works then we know its the code and not the server that is the issue, since we run this exact server setup across all of our shared hosting servers (running PHP in CGI mode) and wordpress without issue.
—

Their logic doesn’t seem to necessarily make sense to me, but interested to hear your thoughts… just going to go ahead and try it to see if it makes a difference or not.

That was my thoughts exactly.

So, they disabled mod_security completely and it was still throwing the 500 error, so that wasn’t the problem. Their response afterwards was this:
—
As we stated before in a few ticket posts back.

In the wp-includes/functions.php at line 928 is the code below which looks like what you are getting above where “/var/www/vhtdocs/userweb1/dead” is the $name and “site-new.php” is the $field_value. I think this is what is outputting this funky header to apache and then apache doesnt like it and gives a 500 error.

@header(“{$name}: {$field_value}”);

If you want, you can remeve this code and see if that fixes it, if it does its something in wordpress and not the server, which is our guess as that is what it is. These values in the list are pulled from either code or the database and its putting them into the HTTP header causing apache to complain.
—

To me, removing the @header doesn’t seem to make sense since this issue should have shown up on one of the other 60 million odd installations of WordPress and I can’t seem to find anything else on this issue.

Coming back after 20 minutes of deeper searching, I just found out that the server is running PHP in CGI mode, so I’m getting them to adjust the server settings to the Apache handler instead based on this thread.

I think so? Or maybe if I’m using W3 Total Cache or a similar caching plugin, none of which I’ve begun to install and tweak yet. I’m going to respond to them and see what they say.

Thanks for keeping up with this and hopefully I’ll be able to have you something more concrete to work with soon.

My host’s response to the mod_security question was:

If it was mod security you would be getting a 403 forbidden message, and it is installed for your protection.

Which means to me that it is installed, but no idea what the setting is at.

I re-created the issue for them and after looking in the logs this was their response:

[Sat Jan 19 08:15:50 2013] [error] [client 68.151.36.45] malformed header from script. Bad header=/var/www/vhtdocs/userweb1/dead: site-new.php, referer: https://kether-interactive.ca/wp-admin/network/site-new.php

Are you using in wordpress a nocache option enabled?

In the wp-includes/functions.php at line 928 is the code below which looks like what you are getting above where “/var/www/vhtdocs/userweb1/dead” is the $name and “site-new.php” is the $field_value. I think this is what is outputting this funky header to apache and then apache doesnt like it and gives a 500 error.

@header(“{$name}: {$field_value}”);

I’m not exactly sure how to respond to this because it seems to me that if this was actually an issue with WordPress Multisite it would’ve been found months or years ago when the very first install procedures were tested… I just don’t know Apache or the tech side of WordPress enough to figure out why Apache is having an issue with the headers being sent. And in particular because the very first time I installed WordPress from a straight vanilla brand new download of 3.5, I was having the issue, I don’t see this being a WordPress thing as much as a random server config setting of some sort.

So, if I go and add a new site to the network, I get the 500 error. If I refresh the page, the page loads and WordPress gives me an error because “Sorry, that site already exists!”.

PHP Version 5.3.20 w/ Apache on Cloud Linux and MySQL 5.1.66