Post Meta

3-4-2008
Date
7:45 pm
Time
161
Views
0
Words
Comments
Comment Feeds

Testing the Slice

I’ve had a slice at SliceHost for longer than I care to admit by now1, I just hadn’t quite gotten around to ever testing and configuring it the way I wanted to.

Well, I finally got around to wiping out whatever I’d been playing with there before and dumping on their stock Debian 4.0 image. After running through their tutorials on setting up Apache2 and PHP, I was good to go.

The base system with my webserver and database running read at about 25 MB of used RAM. Not bad for a fully functional, if barebones, webserver. I’d been worried that, coming from a fully dedicated box with 1 GB of RAM, I would run into a memory bottleneck, but fortunately that didn’t seem it would be a problem.

The next important step was to do some testing. I played around with MySQL, running some basic queries, just to see if it was noticeably laggy after a casual poking. Again, everything looked fine.

The next, and really final, step was to dump a copy of my blog on the slice and see how it ran. After some complaining about the default max_upload_size value in PHP, I got a copy of my database imported using phpMyAdmin and a quick scp -r later and I had an exact copy of my blog setup and ready to go.

All-in-all, it looks like performance is at the very least on-par with the other hosting I’ve used in the past. The performance over DreamHost, where my blog has lived for several months while I really decided where to host it, represents about a 10% improvement3.

I’m still not ready to make the DNS switch, but at least I’ve realized I’m being too paranoid about the memory limits. In the end, the only other reason to stay with my expensive dedicated server is the convenience of Plesk, which scratches my lazy itch perfectly.

If I can get a few scripts hobbled together (in one language or another) to help automate things like vhost and database creation, I may be able to do away with Plesk entirely.

One final problem, and one I’m looking for opinions on, is what to do about email. I’m not planning on dumping DreamHost any time soon4, but I would like to move my email along with my blog if possible.

So who do you use for email? Any problems? Only condition is that they have to offer IMAP

  1. About 6 months, but don’t tell anyone. []
  2. I never expected it to be that different than other distros. []
  3. Going purely by the stats in the footer of my theme. []
  4. I use their massive storage for backups as well. []

Post Meta

Date
6:02 pm
Time
175
Views
442
Words
Comments
Comment Feeds

Missing an Email? It may be Media Temple’s Fault

It started last week when I was trying to sign up for Ron Paul Christmas. For some peculiar reason, I didn’t receive the welcome email. After talking with the site owner, it turned out (mt) was rejecting the email because the email address wordpress@ronpaulchristmas.com didn’t exist on the sending server.

Now, this isn’t particularly unusual. There is no requirement1 that an email address actually exist for a server to send email as if it were from that address. This is especially true from Wordpress blogs, which often send email from wordpress@domain.com accounts on behalf of their owners. Now, since this is only used for outgoing email, in most cases users would never bother setting the email account up. Why would you? You’re never going to be receiving email there2, so what’s the point?

Well, (mt) apparently knows better than you do… For “security reasons”3, their grid service does a “callback” check on every incoming email address. If the server handling mail for domain.com doesn’t recognize that account (such as our wordpress@domain.com example), (mt)’s server will reject the message.

I’ve tried to point out that this kind of behavior can be detrimental, particularly in the age of blogging and web services we now exist in, but the best answer I’ve been able to get out of (mt) is that I should add the sending address to their Mail Protect whitelist. Well great, unless I can add *@* to the whitelist, or at the very least wordpress@*, that’s hardly a viable solution – how do I know the address that’s sending to me if I never get the email?

If you use Media Temple’s grid service4, please contact (mt) immediately and tell them this is an unacceptable situation. I love a lot of aspects of their grid service, but this is clearly not one of them…

  1. In most cases, anyway. []
  2. Except for bounces, should someone put in an invalid email address []
  3. According to the support representative that responded to my ticket. []
  4. Or you want people who do use it to actually receive emails you send to them. []

Post Meta

11-11-2007
Date
2:00 pm
Time
618
Views
352
Words
Comments
Comment Feeds

Plesk Backup Error: Specified file is not accessible

After upgrading my Plesk install past 8.1.1, I encountered a problem with the builtin backup utility. When attempting to create a new backup (either locally or to an FTP repository), I would almost instantly be handed back the error:

Unable to create backup session: Specified file is not accessible

I googled around and found a couple of results, including a support forum that actually had the answer to my problem burried back on the second page.

For whatever reason, Plesk loses the ability to write to its temporary directory, where all backups are held until they are completed (even for FTP destinations). I was easily able to solve this problem by (as root):

chown -R psaadm:psaadm /var/lib/psa/dumps

Note that the original author of the suggestion I used said to chmod 777 the files, but this proved to be unnecessary. I saw that the parent directory was owned by psaadm, and it just made sense that the dumps directory would need to be as well.

In any case, it worked for me. Hope this helps someone…

Post Meta

10-8-2007
Date
12:13 pm
Time
729
Views
177
Words
Comments
Comment Feeds

Take the Apple Challenge

I found Can you tell 128kbps AAC from the original? Take the test! via Digg quite fascinating today.

Since the iTunes Music Store opened, a lot of people have complained about the DRM-infested low-quality music they sold. I’ve often countered that the vast majority of users couldn’t tell one level of high-quality music from another1.

Finally, it looks like this simple experiment has proven my point. If you look at the stats, on roughly every track about half of the people were wrong when they tried to pick the lower-quality and the original tracks apart. If you further take into account that there’s a 50/50 chance you’ll simply guess correctly every time, that means that in fact a huge majority of people don’t have a clue which track is the original – we’ll conservatively say 75% of the overall tested population.

Have you taken the Apple Challenge? Blindfold your ears and see if you can tell the difference between the encoding…

By the way, the site withstood the Digg effect by hosting their audio files on Amazon’s S3 service. Very cool!

  1. I used to have a friend that ripped all his CDs at outrageous bitrates – like 400+ – and claimed he could easily tell the difference. Personally, I think it was just the drugs having rotted his brain, but whatever… []

Post Meta

4-11-2007
Date
10:45 am
Time
194
Views
218
Words
Comments
Comment Feeds