kate's infrastructure

As a compulsive domain buyer, I never felt that compelled to actually move off my old GMail account as my primary email.  It worked well enough, and I got to avoid making any big decisions about which domain would be my favourite.

Still, it seemed like I was missing out a tad.   Gmail is no longer invite only for the tech elite, and prior to that distraction I was using my own domain for email over 10 years ago. 

The irritating thing also was that even though I was using my domain via GMail, it looked bad:

In short, it was time.

I wanted to document the wonderfully simple process I used to make the move, in the hope that someone else might find it useful.

To get started, I assume that you’re already using Google apps for your domain and that you’re up and running at the new location.  If not, you’ll need to do that now.

If you’re like me, you’re primarily using GMail from the website, and even if your mail is downloaded to a few computers, it’s all in cloud.

Now let’s get get GMail ready for the exit:

• Log on to http://gmail.google.com/ with your existing gmail account;
• Go To Settings;
• Click on the tab called Forwarding and POP/IMAP;
• Under the section called POP Download, select the following:
  1. 1. Ensure that Enable POP for all mail (even mail that’s already been downloaded) is checked;
  2. 2. When messages are accessed with POP archive Gmail’s copy;
  3. 3. [Nothing required here – we’ll use Google Apps as the mail client]
• Click Save Changes 

This means that we’ve made all of your GMail messages available for download from someplace else.  It also means that anything downloaded will be Archived by GMail – so you’ll know what’s been transferred.

And the next step is to tell your domain to transfer your mail over from GMail:

• Log on to your new google apps mail account at https://mail.google.com/a/yourdomain.com or similar;
• Go To Settings;
• Click on the tab called Accounts;
• Under the section called Get mail from other accounts, select the following:
1. Click Add a mail account you own;
2. A popup window will appear;
3. Enter your GMail address at the prompt, Click Next;
4. Enter your GMail credentials;
5. The POP server should default to smtp.gmail.com – this doesn’t need to change;
6. Check the box marked Always use a secure connection…;
7. Check the box marked Label incoming messages (this will make things MUCH easier later);
8. I didn’t check the box for archiving incoming messages, but you might prefer to;
9. Click Add Account; 
• Click Save Changes;
• Your mail in your Gmail account should begin downloading (oldest first) immediately – the rate doesn’t exceed 200 mails an hour, so expect the transfer to take a while.  Mine took over a week;
• Remember we chose to archive mail in Gmail as it’s transferred, meaning you should start to see the quantity of mail in that account reduce;

You can choose to remove this link later,  but I found it easiest just to leave it as is.

Other stuff I learned:

• Importing and Exporting mail filters was easy and totally worthwhile;
• Contacts were perfectly exported, imported and useful;
• GTalk chat logs aren’t transferred;
• I really should have cleaned up my gmail inbox a bit first – preferably by more frequently deleting or archiving the mail I no longer need;

My dreams of the vanity domain and google apps have successfully been realised.

Frustrated by my inability to get Microsoft Word to do some pretty simple formatting I wanted, I turned to the internet for help.  Turns out what I wanted was pretty hard to devise a sensible search string for*, so I eventually worked it out on my own.

This is as much for my future reference as yours, but:

Stop a pagebreak occurring in a table row

• Table > Select > Table
• Table > Table Properties… [a table properties dialog is launched]
• Select the tab called Row
• Check the box next to Allow row to break across pages
• Click OK

Stop a table running over a pagebreak

• Table > Select > Table
• Format > Paragraph… [a paragraph properties dialog is launched]
• Select the tab called Line and Page Breaks
• Check the box marked Keep with Next
• Click OK

It’s worth also noting that this isn’t table specific, so you can avoid oddly placed pagebreaks by selecting any passage of text this way.

Create multiple references to a single footnote

• Create a footnote as usual (Insert > Reference > Footnote…);
• Then Insert > Reference > Cross-Reference;
• From the dropdown called Reference Type, select Footnote;
• From the dropdown called Insert Reference to, select Footnote Number;
• From the textarea called For which footnote, select whichever note you’d like to reference;
• Click Insert;

This was useful as I didn’t want the footer of every page to be flooded by repeating the same disclaimer.  This way I can place the original footnote reference in one prominent location in the document and refer to it again throughout.

*Actually I had the perfect search result, but it was a link to experts-exchange.com, so as I non-member I fail.

Ignoring the smarts that may exist in off-the-shelf backup software these days, I’ve decided to include some info on how I backup my laptops*.  I’m also without a fancy one-touch hardware based backup solution, but I was equipped with:

• 1 x 320gb drive
• 1 x Vista Laptop, 80gb total drive space
• 1 x XP laptop, 80gb total drive space
• A general desire to make things as easy, painless and low-tech as possible.

Step 1 – Pre-requisites

To begin, I downloaded Robocopy from the Microsoft website, note that this is a part of the Windows 2003 resource kit and is required on the XP machine only.  Vista is already that smart.  Robocopy is a Robust File Copy utility which we’ll use to do all the hard work for us via batch files.  Watch out for rogue copies of software pretending to be robocopy, downloading from the Microsoft website is the best bet here. 

Next, I need to plug in my Backup drive to my computer.  If your computer has USB 2.0 as well as 1.1 ports, be sure to choose the 2.0 one to ensure the fastest transfers.  Windows should let you know if you’ve made the wrong choice.  When you’ve plugged in the drive, open Windows explorer and take a note of the drive letter.

Step 2 – Scripting magic

Here’s a simplified version of what I run:

echo “Starting backup process \n”

ROBOCOPY c:\ g:\firstLaptop /e /zb /R:3 /MIR

echo “Completed backup process.”

I told you Robocopy was worth the 11mb download. 

In pseudocode that would look like this:

[Run Robocopy.exe] [Copy from] [Copy to] [use these options/flags for Robocopy] 

Let’s look at the flags in use:

• /e  = Copy subdirectories, including empty ones
• /zb = Use restartable mode
• /R:3  = Number of retries if a file copy fails (default is 1 million!)
• /MIR = Mirror a complete directory tree 

You’ll see that these options are all about maintaining a mirror of my laptops and taking the intelligence out of the process.  The first time I run it, it takes forever.  The next few times its appreciably faster as it performs a file compare. 

Step 3 – Room for improvement

This pretty much meets my needs as neither of my computers have anything mission critical aside from photos or mp3s.  Everything else is either online or recreatable.  Primarily I’d put in a timestamp at the start/stop of the job and use some of the additional features of robocopy to make it prettier and easier.

* This in now way addresses the need for multiple off-site copies etc of your data.  I may address this at a later time.

I wanted to report a weird issue with a website I was visiting.  I thought it would be useful to include info about my platform, but really there’s a heap of different values that might be of interest to someone who’s trying to debug an issue.  I figured the most useful and most “user-trying-not-to-interpret-it-themselves” version was to send the HTTP user-agent header value. 

Overcome with laziness, naturally the internet has the answer to finding this this: http://www.ericgiguere.com/tools/http-header-viewer.html

It’s fact that xkcd is a damn fine web comic.  I’d be happy to have just linked to that.  Because I also need to be entertained to learn, I choose to present you with xkcd’s map of the internet in all it’s comicy goodness.

MetaFilter just also sent me in the direction of a much more sensible version, which is an interesting reference.

A relative had a child recently, and I thought it was a very ‘now’ thing to do to register them a domain name.  It might not be for much more than baby photos now but it might be a useful myspace redirection when they’re an angsty teen.  Completely infatuated with my new web host I thought I’d kick the domain registration their way.  A few clicks and I’d confirmed the availability of the new domain and registered it.  Awesomely easy and I felt like a technical genius. 

As my webhost of choice acts as both a host and a domain registrar, I had to run through the final step of setting it up to be actually hosted (2 month old children aren’t awesome at doing such things themselves).  It failed immediately, which led me to believe I’d done something wrong.

I checked their support wiki, I’d appear to have followed the steps.  Really there’s no rocket science involved, so I don’t see how I could have screwed it up.  That being said, it hadn’t actually worked, so evidence suggested that I had in fact screwed something up.

Next was time to eliminate some of the obvious stuff.  Unfortunately I had a pretty short list:

1. Try again.  Nope, still doesn’t work.
2. Try again using a different option (e.g. Park the domain).  Nope, still doesn’t work.
3. Check that the domain name has propogated.  Yes, whois records show my registration and I can reach it from other sites on the interweb.  All a-ok.
4. Um, Try Again.  Still doesn’t work.
5. Raise a support request.  Try to not sound like a lunatic.

(Yes, I should be in testing)

A quick back-and-forth with the support team and I learn that I can’t add my newly registered domain to my host because it’s already registered under an expired account there.  No, that wasn’t in my list above of random tasks to try.  Seriously though, even if it was in someone’s account they didn’t hold the registration for the domain itself.  Obviously it wouldn’t have let me register a domain that belonged to someone else, and I did a WHOIS from a different registrar to start with, so as far as the internet was concerned, this was up for grabs.

I tried to plead my case, as this sounds a little crazy, and my request has been moved to a different support queue.  It got me thinking though, aside from my receipt for payment, I have nothing to refute the history of the domain.  I didn’t even think to check this.  What if it was chock full of unsavoury content in a previous incarnation?  How can I find out what has gone before? 

1. Google Cache: I decided that this would show me nothing so I didn’t check it first.  I generally use the cache to check out sites that are offline when I need them. 

2. WayBack Machine: The Internet Archive is the most awesome at this sort of venture.  Ultimately you either love or hate that it exists.   There’s always stuff you wish wasn’t archived, but it usually pretty useful.  No matches for my site.

3. Search for WHOIS history tools:  I guess this is all I really wanted, I just wanted to confirm that it had been registered before.  I stumbled upon domaintools.com and it looked perfect.  Unfortunately the front page did little service to a skim reader such as myself and I came away with the conclusion that additional funds were required.  A little frantic searching later and I found their ultra awesome subdomain: http://whois.domaintools.com/.  Running a WHOIS from there not only returns the WHOIS record, but gives me some added extras including the host, web server software and history.  Perfect.

4. Google Cache: I did eventually check google for completeness.  It did return me results indicating my domain of choice had come up on lists as about to expire back in March.  That was enough info for me to use in the first place to confirm that it had been registered previously and that the registration had long since expired. 

Conclusion: The unexpected will always arise, the simplest solution is best and bookmark whois.domaintools.com.

Check out the NetCraft web server survey and news site: http://news.netcraft.com/

It’s pretty straight forward and contains some interesting data about web server trends and usage.  It’s particularly interesting to see major organisation downtime (Blogger, Cisco) as well as some of the shift away from Apache.  Microsoft’s IIS and the not so Microsoft lighttpd are emerging as interesting players.  I think I might need to take another look at these.

Storage illustrated through russian dolls: http://www.artlebedev.com/everything/matryoshkus/

Excellent, except for ending with Terabyte.  Must future proof gimmicks.

A reminder of what’s already in place, and maybe the soon to be common:

• Kilobyte (KB) - 1024 Bytes
• Megabyte (MB) - 1024 KB
• Gigabyte (GB) - 1024 MB
• Terabyte (TB) - 1024 GB
• Petabyte (PT) – 1024 TB
• Exabyte (EB) – 1024 PT
• Zettabyte (ZB) – 1024 EB
• Yottabyte (YB) – 1024 ZB

I remember having under a GB storage on my home computer, and in 2007 a TB is now in reach for home users.  Even Mac users. Crazy times.

I’ve not used this, but if I don’t record the link someplace I’m bound to need it.  This looks good: http://home.eunet.no/pnordahl/ntpasswd/

Though it says NTPassword, it words right through to Vista.