Steve Jamesson » servers

Gmail troubles? Calm down, have some dip.

Steve — Wed, 25 Feb 2009 18:40:15 +0000

I heard lots of people complaining about recent Gmail issues over IMAP. I mostly use my gmail account as an archive, so all my emails end up in there. It works really well, for me at least.

But there’s an issue today, where I can’t seem to move my messages into the All Mail (archive) folder. It’s annoying and is keeping my inbox too crowded.

But do I pay for Gmail? No! It’s free. And my storage space is ever increasing.

So, if you’re having trouble, just be patient. Google will fix it up. Don’t forget, the service is FREE!

What the hell is web site hosting, anyway?

Steve — Wed, 08 Oct 2008 20:59:59 +0000

I answer this question quite often, also, so hopefully this article will help to clarify what website hosting is. Website hosting (aka hosting) is much different than a domain name, though they’re very closely related.

As I discuss in the above article, a domain name is simply a name, registered via an authorized registrar. An A Record beneath the domain points to an IP address, and that IP address routes to some computer located somewhere in the world and connected to the internet.

This computer is your host. Also known as a server or web server, this machine is responsible for storing and serving your website files to people who visit your website.

Host computers (aka hosts) vary greatly in performance, speed, and versatility. Most websites’ hosts run Unix-based operating systems. Others may Windows and IIS, which stands for Internet Information Systems.

Windows based servers

A Windows server is simply a computer that runs a flavor of Microsoft Windows, ranging from Windows 98 to XP to Vista. The operating system (Windows) isn’t the one doing most of the work, though. IIS is actually responsible for accepting requests for a web page and serving responses for that web page. IIS is a web server software.

Unix based servers

On the other hand, the majority of host machines are Unix-based, and most often run a Linux operating system. Linux comes in many flavors, including some of these popular ones:

There of course more, but these operating systems are very commonly encountered.

Linux-based servers still run a core that’s based on an old (but very solid) operating system (OS) called Unix. You may encounter the term **nix* OS. This term applies to any OS that’s based on Unix, including Mac OS X, Linux, and others.

Most *nix servers run web server software called Apache. Apache is by far the most widely used web server software. Like Microsoft’s IIS, Apache is the one responsible for responding to requests for web pages.

How to get hosting

There are many companies that offer hosting, and if you’re searching, you’ll probably easily come across thousands. Many domain registrars additionally offer hosting services. Lots of these companies offer SSL certificates, email solutions, and other services.

Using one company to manage your domain names, hosting accounts, certificates, and other products is a great way to keep everything centralized and easy to find. Of course, there’s the risk that the company’s site goes down or goes out of business, but I think the benefit of keeping everything centralized is worth that risk.

While you’re shopping for a host, I highly recommend you check potential hosts on Host Jury, a great host review site! Check the host’s rating on Host Jury before you sign up!

Find out which version of Apache is running on unix/*nix systems

Steve — Tue, 07 Oct 2008 20:57:21 +0000

To find out which version of Apache you’re running on a *nix system, you can run the following command. This will only work if you have permission to run the script in the first place, so you should probably login as root or use sudo, if you can. If this isn’t possible, then you should probably just contact your hosting provider or just look around in your hosting control panel.

steve$ httpd -v

or use this

steve$ httpd -V

Project Pier upload error: Failed to create folder

Steve — Tue, 07 Oct 2008 20:55:55 +0000

I’m evaluating a collaboration/project management web application called Project Pier. So far I like it, but I got a weird upload error when I was testing file uploads. The error read:

1	Failed to create folder: /var/...xxxx.xxx/httpdocs/uploads/

My first reaction was of course to check permissions on the /upload directory. It was already set to 777, so I was a bit confused.

A little research led me to this page on Project Pier’s forums. It of course mentioned a solution posted on another page on the same forums. This solution involved replacing the default function force_mkdir() with a new one.

I modified the file

1	/var/...xxxxx.xxx/httpdocs/environment/functions/files.php

and replaced the original force_mkdir() with the code found directly on the second forum page. It worked! Some people reported it didn’t work for them, but it worked for me under Linux, RedHat, Apache 2.x, PHP 5.1.x.

Here’s martinlang’s replacement code, reposted for your convenience. This is identical to the code he posted at this page.

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28

function force_mkdir($path, $chmod = null) {
if(is_dir($path)) return true;
$forced_path = '';
$real_path = str_replace($_SERVER["DOCUMENT_ROOT"], '', $path);
$forced_path = $path != $real_path ? $_SERVER["DOCUMENT_ROOT"] : '';

$real_path = str_replace('\', '/', $real_path);
$parts = explode('/', $real_path);

foreach($parts as $part) {
if($forced_path == '') {
$start = substr(__FILE__, 0, 1) == '/' ? '/' : '';
$forced_path = $start . $part;
} else {
$forced_path .= '/' . $part;
} // if

if(!is_dir($forced_path) && $forced_path!='/') {
if(!is_null($chmod)) {
if(!mkdir($forced_path)) return false;
} else {
if(!mkdir($forced_path, $chmod)) return false;
} // if
} // if
} // foreach

return true;
} // force_mkdir

Hope this helps someone out!

What the hell is a domain name, anyway?

Steve — Tue, 07 Oct 2008 20:55:04 +0000

A domain name is simply a name ending with a TLD. TLD stands for Top Level Domain. Some examples:

stevejamesson.com
irs.gov
google.com
abtech.edu
cakephp.org

The TLD is the .com, .org, .net, etc. portion of the domain name.

Domain registrars

In order to have a website, you’re gonna need a domain name. You can register a domain name with a whole bunch of different registrars. A registrar is simply a company that’s authorized to register domain names. There are many registrars, including these popular ones.

There are of course many other sites that will offer to register a domain name for you.

How much is it?

The general cost of a domain name is typically no more than about $20 per year, so if you’re currently paying more than that, then you might want to consider transferring your domain name to a different registrar to save some money.

Think alias

A domain name is kinda like an alias. Think of it as speed dial on your phone. Behind every domain name is what’s called an A Record. An A Record points a domain name at an IP address. An IP address is kind of like a phone number, except it’s for computers.

So when you use speed dial number 3, you call a phone number like (111) 222-3333. Well the same thing kinda happens whenever you type in google.com or another domain name.

Your web browser looks up the domain name via a name server, generally provided by your Internet Service Provider (ISP)
The name server tells your web browser that it points to some IP address (i.e. 11.22.33.44)
The computer located at that IP address takes it from there, and you’re shown a website

So it’s kinda like a speed dial. The technology of how all this stuff works is pretty heavy, so I won’t go into those details here.

Is that it?

Yeah, that’s pretty much it. To summarize:

A domain name is simply a name with a .com, .net, .org or other TLD.
Behind a domain name is an A Record, which points the domain name to an IP Address, which is a numeric address for one computer.
That computer takes over from here, showing you a website

Using Gmail IMAP with Apple/Mac Mail application (Mail.app)

Steve — Sat, 04 Oct 2008 20:49:07 +0000

If you’re using Mac Mail and have a Gmail account, you should probably follow a few simples steps to get the Gmail account integrated into Mac Mail. This article will go through how to setup Gmail in Mac Mail as an IMAP account.

First you should understand the difference between IMAP and POP3 mail accounts. When you set up your email account as a POP3 account, you’ll be downloading full copies of the messages from the server, and in many cases after download, the messages will be removed from the server. This approach is best if you really only use one computer, which can hold all your mail.

But if you’re like me, then you may work from multiple computers, from multiple locations, and you may even use your phone for email. Enter IMAP.

IMAP stands for Internet Message Access Protocol. That makes sense, huh? The real beauty of IMAP is that you can access mail from multiple locations, and as opposed to POP3, your inbox and other mail folders will all be synchronized. That’s the main benefit of using IMAP. Rather than have multiple copies of your email messages (as with POP3) across multiple computers or devices, using IMAP is a great way to keep your email account synched across all your devices.

After creating the mail account

After you create the new Gmail account in Mac Mail.app, make sure to set the IMAP prefix. You can find this field under Mail > Preferences. Choose the Gmail account, and then visit the Advanced tab. You’ll see a field titled IMAP Path Prefix. Enter [Gmail]. (Yours may be [Gmail account] or something similar). Then close mail preferences and click Save when it asks.

Now you’ll see the account in the left column of Mail.app’s interface. You’ll also see several folders, including the common folders

Drafts
Sent Messages
Deleted Messages
Spam

You next need to tell Mac Mail.app to use these folders for its common tasks.

Select the Gmail account’s Drafts folder
Click the top menu Mailbox > Use This Mailbox For > Drafts

After this you’ll notice that the folder disappears from the Gmail section; it’s now located up top with the rest of your mail accounts. This means that Mac Mail is treating the folder as your primary Drafts folder, at least for your Gmail account. So if you’re writing an email from your Gmail account and you save a draft, then you’ll be able to access it via this folder. Additionally if you visit the Gmail webmail interface, you’ll notice that the draft is also located in Gmail’s default Drafts folder. Cool! Seems to be working.

Now repeat the above steps for the other folders

Sent Messages > Sent
Deleted Messages > Trash
Spam > Junk

That should do it! Your Gmail account is now integrated nicely with Mac Mail.app.

If you’re still having trouble or want to look a bit deeper into it, I’d suggest starting at this very helpful post on 5ThirtyOne.

Reset Mac OS X network configuration to factory default; Airport interface disappeared

Steve — Wed, 01 Oct 2008 17:43:38 +0000

We used Apple’s Migration Assistant Application to migrate my brother’s files and user account and stuff over to a new machine. The new machine has airport.

After the migration (and a few changes to our simple local home network), the Airport interface in System Preferences simply wasn’t there. Instead all we had was Ethernet 2. The new machine (an iMac G5) only has one ethernet interface, so Ethernet 2 made no sense at all!

I renamed the Ethernet 2 interface to Airport, turned it off and back on again. AFter this, Airport wouldn’t connect to the wireless signal, and the machine automatically renamed it back to Ethernet 2. Frustrating. Tried this about 5 times, and still all I had was Ethernet 2.

So, a bit of research led us to attempt to reset the network settings back to factory default. It was pretty easy to do this. All we had to do was to remove/rename a preference file and restart the machine. On boot the system will recreate the preference file from its default sources.

To do this, I renamed the file:

    /Library/Preferences/SystemConfiguration/preferences.plist

to:

    /Library/Preferences/SystemConfiguration/preferences.plist.bak

and restarted the machine. It worked!

The reason I choose to rename it instead of deleting it is just in case it didn’t work, I could at least put the old back. It’s always best to keep a backup before removing files, especially if you’re not really sure what the hell the file even does in the first place!

What the hell is a blog, anyway!?

Steve — Tue, 30 Sep 2008 17:42:06 +0000

A client recently requested a quote for a new website. The site required at least five different sections, where the site owner would have the ability to administer the site. The proposal requester mentioned that maybe a blog would be sufficient for the first section. He also suggested that maybe a blog would be sufficient for the second section. For the third section… yup, another blog!

Now, this doesn’t seem right at all, huh?!

I believe there’s a lot of confusion when it comes to the term blog. This term gets thrown around all day long, by people who know not what a blog really is.

Simply put, a blog is merely a collection of articles written by one person, generally arranged by date, online and accessible via a web browser. Lots of blogs also use some means of taxonomy, or in other words, they use some form of categorization or classification of articles.

The term originally comes from web log, which refers to a log of articles, diary entries, or even notes organized by date. After a while the two words were combined, and hence our buzz word blog.

A blog is comprised of posts

Each entry in a blog is a post. A post could be in article format. So one blog post isn’t a blog. It’s just a post within a blog.

A blog is targeted

Most blogs (at least successful ones) are targeted to a specific audience. Many bloggers are also web designers or developers, and lots of them will have blogs about–surprise, surprise!–web design and web development.

Other bloggers know nothing about web design or development, but they do know lots about whatever field they’re in. Marketers, researchers, salesmen, schemers, entrepreneurs, and thousands of others might blog about their specific field. They may also have huge audiences

But the key here is that blogs are generally targeted to a specific market or audience.

A blog isn’t an ecommerce site

Lots of blogs have ads plastered all over the blog. This is a way for a blogger to earn some income from sharing her valuable information. If her readers appreciate any of the ads and click on them, the blogger may get some income out of it.

But a blog doesn’t have to sell anything. It doesn’t have to display ads. Again please see the above definition of a blog.

Summary

Stop referring to each blog post you make as a blog; the whole thing is the blog, not just the one post. And if your site needs five different blogs within it, then maybe you’d better rethink the organization of your site. Maybe you just need a content management system (aka CMS)…

Microsoft killed our definition of FTP

Steve — Mon, 29 Sep 2008 17:40:56 +0000

When dealing with so many clients as I have over the years, and lots of them being primarily Windows users, I’ve come to a point where I can’t stand hearing incorrect definitions of FTP.

It seems like Microsoft has incorrectly impressed the idea of FTP in so many of its users’ heads. Many clients and friends ask me about having an FTP area. An FTP area doesn’t really exist!

FTP stands for File Transfer Protocol, which means it’s a way of transferring files back and forth. FTP is actually just a program, like everything else that runs on a computer.

Compare FTP to HTTP. Many are familiar with http://www.example.com. HTTP stands for Hyper Text Transfer Protocol. What this means is that it’s a way of transferring hypertext data. Hypertext data simply put is just HTML. It’s not that complicated, but there are lots of intricacies to it.

Here’s a quick example of some simple HTML (HypterText Markup Language):

A few notes

Here's a few notes that I gathered at a lecture I recently attended.

    Germans make good German beer
    Hops were originally added to beer as a preservative
    Wheat beer is best unfiltered
    Use an anchor to link to google

Take a minute to look it over. It's quite simple. The

is just a heading, same as you'd have in a Microsoft Word document or other word processing document. The `3` denotes that it's a level three heading. You could also have
,
,
and so on up to
.

The
simply encloses a paragraph.

The
represents an unordered list. And the
s represent items within that unordered list.

I've also included an tag, which stands for anchor. This is how you'd create a link to another website.

So hypertext really just means metadata, in this sense. It's data that describes the text in which it's embedded. So the hypertext
and
says that its enclosed text is a paragraph. Same with the unordered list and list items. Pretty simple enough to grasp, huh?

Now onto FTP. Again, FTP means File Transfer Protocol. And again, this means that it's a way of transferring files. It's a program that allows a user to `put` and `get` files to and from one computer to another.

Hundreds of times I've heard Windows users use the term FTP Area for our files. Well, you can't really have an FTP area, since it's only a way of transferring files to and fro. You can have a file area. You may decide to access it using FTP, if you'd prefer. There are other ways of accessing these files other than using FTP.

FTP is pretty rudimentary, but it does allow the user to do everything she needs to do when it comes to transferring files.

So please, understand that an FTP Area doesn't, and can't, exist. It's that simple.

Ubercart using NaviGate/Authorize.net payment gateway

Steve — Fri, 26 Sep 2008 17:38:43 +0000

I had some trouble getting Ubercart’s Authorize.net module working with a different gateway named NaviGate, provided by MerchantPlus. The NaviGate gateway supposedly works just like an Authorize.net gateway, with the same variable names and very similar API, so we’d hoped we wouldn’t run into many issues. But there was one that caused a decent amount of frustration.

The issue was related to the credit card’s expiration date every time an order was attempted. The Ubercart error message was:

Payment failed: Credit card payment declined: The credit card expiration date is invalid.

Now I was using my own personal credit card, and I’m a friggin web developer, so I’m pretty sure I had input the damn expiration date correctly! But nonetheless, I kept getting the same error messages.

After a bit of time and as the debug mindset sank deeper in, I checked the logs once again. Noticing an error logged from the uc_payment module, I checked over the string that was cURLed over to the gateway. The expiration date was sent as 1%2F2009. Ok, that seems right. The %2F of course is just a slash, no worries there.

I kept thinking about the issue, and eventually I decided to try something I’d previously written off. I though maybe, just maybe, the expiration month and must be two digits, and the year four digits. The cURL string above was using just one digit for the month.

I found the code that generates the select box in uc_store.module. I didn’t expect it to be there, but oh well. Around line 2276 (in my copy of the file at least) I found the array that’s used to build the expiry date select box.

/**
     * Create a month select box for a form.
     */
    function uc_select_month($title = NULL, $default = NULL) {
      $select = array(
        '#type' => 'select',
        '#title' => (is_null($title) ? t('Month') : $title),
        '#options' => array(1 => t('01 - January'), 2 => t('02 - February'), 3 => t('03 - March'),
                         4 => t('04 - April'), 5 => t('05 - May'), 6 => t('06 - June'),
                         7 => t('07 - July'), 8 => t('08 - August'), 9 => t('09 - September'),
                         10 => t('10 - October'), 11 => t('11 - November'), 12 => t('12 - December')),
        '#default_value' => (is_null($default) ? 1 : $default),
      );

  return $select;
}


So I simply changed the keys to be strings at two digits each:

...
'#options' => array(
    '01' => t('01 - January'), 
    '02' => t('02 - February'), 
    '03' => t('03 - March'),
    '04' => t('04 - April'), 
    '05' => t('05 - May'), 
    '06' => t('06 - June'),
    '07' => t('07 - July'), 
    '08' => t('08 - August'), 
    '09' => t('09 - September'),
    '10' => t('10 - October'), 
    '11' => t('11 - November'), 
    '12' => t('12 - December')),
'#default_value' => (is_null($default) ? '01' : $default),
...

I also cleaned up the alignment to make it a bit easier to read, even though I'll probably never look at it again. Guess I'm kind of anal coder, but that's a good thing, right? Notice that the default value also was changed to '01', rather than 1.

This seemed to fix the issue. Live transactions worked right away!

Remove CalDav subscribed calendar from iCal, calendar won’t delete

Steve — Thu, 25 Sep 2008 17:35:37 +0000

I used to use Leopard Server for my office. We all had accounts on there, and we used iCal and CalDav so we could all subscribe to a few common calendars. It never worked perfectly. But now we don’t use Leopard Server anymore. It just didn’t really help us out very much.

In my iCal I’d subscribed to a master calendar from the server. I’ve deleted that account hundreds of times, and every time I restart iCal, it show back up! And of course it complains that the connection can’t be found. No kidding the connection can’t be found; the server’s not there anymore!

The fix

You can remove the iCal managed preferences file. It’s in:

~/Library/Preferences/com.apple.iCal.managed.plist

Just delete it and restart iCal. I had to restart iCal twice before I noticed the ghost calendar disappear. But this should take care of it.

Clear or Flush DNS Cache on Mac OS X Leopard

Steve — Mon, 22 Sep 2008 17:21:32 +0000

Sometimes it’s necessary to clear DNS cache on your local machine. If lookups are sluggish, it’s most likely your ISP, but it’s worth trying to clear cache on the local machine.

To clear cache on Mac OS X, open a terminal window and type the following command:

dscacheutil -flushcache

Then click enter. You won’t get any output from running dscacheutil but the cache has been cleared.

Try some more lookups, but you may need to restart the machine or logout/login to have the cache flushed.

Unix get full size of directory folder

Steve — Wed, 18 Jun 2008 17:04:16 +0000

This issue has plagued me for awhile, but I never figured out the command to actually have unix print out the size of an entire directory. I came across the need handfuls of times over the years, but today I finally needed to actually learn this one.

So a bit of research led me to the following command:

steve$ ls
dir1  dir2  dir3
steve$ du -sh dir1
2.4M  dir1

Finally I can get the full file size of an entire directory via a shell. Hopefully I’ll remember that I posted this so message, for the next time I forget about this command.

By the way, about the arguments above: -s tells the shell I want size, and -h tells it I want it human readable.

How to check Plesk mail logs

Steve — Wed, 14 May 2008 16:59:03 +0000

Plesk uses qmail. You can check your mail logs here:

/var/log/maillog

I use this command to follow the maillog as it runs:

tail -f /var/log/maillog

You can use Control+C to get out.

Plesk, Email Piping/Email Pipe (osTicket / Trellis)

Steve — Tue, 13 May 2008 16:52:45 +0000

I was trying to get email piping working on a RHEL server that runs Plesk (8.3), and I was having no luck. I just wanted to test out both the osTicket and Trellis support ticketing apps, but I had the toughest time getting either to work well on Plesk.

I did finally get it working for osTicket, both locally and remotely.

You’ll need to modify your /var/qmail/mailnames/[domain]/[address]/.qmail file. By default, mine looked like this:

| true
./Maildir/

To get piping working, all I did was add the following:

|true
|/path/to/pipe/handler.file.php
./Maildir/

There were some gotchas:

Make sure the handler file has executable permissions, and make sure that its full path has executable permissions.

So if your file is in
```
/my/great/file.php
```
then my, great, and file.php will need executable permissions.
Some people said quotes were necessary around the piped command (i.e. "|/path/to/pipe/handler.file.php") but I’m not using any, so I’m not sure that’s your biggest problem.

Getting it working with osTicket took quite some time, but it seems to be working great now! I did find lots of info about how easily cPanel handles this type of thing; they offer a nice text input for pipe commands via the admin interface–how convenient!

Here’s a few other resources. They weren’t incredibly helpful, but they did eventually lead me in the right direction: