I think we all know by now that 500 means an Internal Server Error (given how much I’ve had to mention it recently). Usually it’s caused by the server you’re on being annoying and not working properly. In those cases it’s nothing you can do, so you need to get over it quickly. However, most other times it appears because the server doesn’t seem to like the way you lay out one of your system files, and then decides not to try and serve anything. Annoying!

A few months ago, Donncha shared with me his tactic of getting rid of site scraper bots for good with a little bait and trap. It all worked well back on GoDaddy, but ever since moving to Blacknight all of my .htaccess allow and deny records have thrown up 500 errors… Removing them solved the 500 problem, but alas not the scraper problem.

So, this morning, checking my quite large inbox of new mails (90% of which are new comment notifications) – I notice a little e-mail from Apache. Not the pizza place, but an Apache web server. Another site scraper banned. Didn’t phase me all too much, given how many of them had been caught over time. Onto Twitter, and I notice a little @reply from Reflex, sent Friday evening:

@TheChrisD What up with your site?

My first reaction was WTH? But as I open my admin panel, sure enough there’s a 500 error there. A little digging around later shows that my trap was doing its job in catching the spider and adding it’s deny order to my .htaccess.

Only problem is, not only did I remove the allow and deny base syntax, but either way it caused a 500 error no matter what! So, since Friday my site had been down all because of my attempts to stop my site being scraped…

The whole 500 issue is one of only two things that do annoy me quite a bit about Blacknight. The other being the problems with open_basedir, which prevent WordPress from auto-updating itself because it can’t seem to locate the installation folder…

Either way, once I get the weekend’s backlog of feeds out of the way, I can shoot off a little e-mail to the BN support team with my concerns. Unlike GoDaddy, they’ll get back to me on average within the hour – absolutely thrashing GoDaddy’s previous record of… 29 hours.

PS: Before I forget, GoDaddy tried to auto-renew my hosting with them over the weekend, which thankfully failed Had they got away with renewing without any notification that they were going to charge my card, I would have reported them to my bank for Internet fraud!

PPS: Looks like Wordbook fucked up when I tried to post this… Oh well, Facebook will import the “note” eventually anyway.

Anyway, this whole thing started after Alexia moved from the breaking-every-half-hour-crappy BlueHost to Blacknight. I had always wanted to have my site(s) hosted on an Irish server, but up until recently, it all seemed overpriced in comparison to what I could get on the huge American hosts. Either pay €4 a month for 1GB of space and 10GB of bandwidth, or pay $4.20 a month for 5GB space and 250GB bandwidth? I think the answer is obvious!

Blacknight had always had a special blog hosting package available that would have suited my needs (OK, well maybe not my disk space needs), but the only problem with it was that you could only have a single domain on it, and no sub-domains at all. Given I use stuff.thechrisd.com as my hotlinking domain, it didn’t cut it. Recently though, they re-launched their service with new packages and site and other lark. The new most basic package was roughly the same price (€4.95/month) that was I was currently paying to GoDaddy ($6.99/month), and still came with more than what I’d need: 10GB space, 100GB bandwidth and the usual extras. Not exactly “Minimus”!

Friday

Right, so Friday afternoon, I bought myself a nice little one-year Minimus hosting package from Blacknight Things were looking up as I eventually uploaded my website backup to them, changed the configs, rerouted the domain DNS and everything. The RPWS site looked good, but I couldn’t really say the same for my site… Turns out I had made some changes since my backup from the end of July. A LOT of changes. Which led at first to some 500 errors, then to bigger and worse problems… I was beginning to think that this was going to turn into a hosting hell…

Now before you start thinking that I was all alone with a broken website, I have to commend the support services at Blacknight. There was always a very fast response from the team to my queries, and although I may have whined a bit about something, and I may have been a nuisance, their team always came through to attempt to help out. Seriously, I cannot stress how great their support team is.

Back to Friday, I was getting pressed for time, and couldn’t seem to fix the problems, so I had the DNS changed back to GoDaddy, with plans to re-attempt to transfer sometime next week. Which turned out to be today!

Monday

The first step I wanted to get done was to move the databases to Blacknight and get them setup and running before moving all the files over, so that I wouldn’t lose any Statpress logs, new comments (like what almost happen to one of Reflex’s comments on Friday!), and other things. All seemed to go according to plan, until the database connection errors started…

Onto the support team again just to make sure I wasn’t screwing up with the external options. Blacknight have the handy feature of allowing external hosts to access your databases, so I setup this option specifically for the move. Well, it seems that GoDaddy were having none of that, since the BN MySQL servers were getting no traffic from my GoDaddy site… Epic fail. Hence the few hours earlier where I had a filler page saying “Gimme a few minutes while I sort out databases”. Well, that kinda turned into a few hours…

I sent off a support mail to GoDaddy asking whether they allowed me to access other databases from my server, and whether they could turn it on or not for me. Realising though, that they’d take ages to get back to me, I went ahead and continued with making a new, more up-to-date backup of my site and uploading it to BN. Several hours later and everything is up to BN – and still no response from GoDaddy support!

Finishing Touches

There wasn’t really much else to do after uploading everything. Gallery was giving me a few errors on my site at first, and then was just giving notices, although I do have to blame that on the fact that I had left debug mode on (after the errors on Friday!), but that was soon sorted. Everything else looked fine, and as of now seems to work fine. All in all, today’s move was a LOT smoother than Friday’s!

Overall Opinions

So, based on this move, I’ve changed my mindset regarding a few things:

• Switching to a more local host is MUCH better, even if it may cost more or if you don’t get as good a deal as what you had before (judging by the fact it took only half an hour to upload all my files, while it took 3 hours to download them from GoDaddy!)
• Try not to forget to turn debug mode off…
• DNS can resolve either very fast or very slowly, depending on which you don’t want. When I wanted it to take a while earlier, it happened quickly…
• Remember that your FTP client can hide your .htaccess on you…
• GoDaddy support sucks balls. Seriously big, chocolate salty balls…
• Michele Neylon is one of the best CEOs there is
• The DCU FTP proxy fails miserably when used with my FTP client

So yea, that’s about it. When you read this, try and remember it’s powered by an Irish web hosting company, that perform a lot better than the big, over-the-top American hosting providers.

PS: The RPWS site is still as of now using the GoDaddy services. I still haven’t moved it yet, but then again it’s not the number one priority here!

Anyway, it was about 3pm in the afternoon yesterday when this started. I did a little poking around and found out that FTP was still working, but no HTTP traffic was getting through. By half 3, I had fired off an e-mail to my hosting support guys asking if there was a problem with the server I’m on. Only problem with my support guys is that they say "Please wait at least 24 hours for a response". Which doesn’t help for urgent stuff. The phone support might have helped a lot sooner, but I’d rather not have to pay huge over-inflated rates to make international calls just to get my $4 a month website back up and running quickly. The call would cost more than what I pay to keep this hosted…

Anyway, by the early evening, I had given up trying to find a solution myself, and went to play some games. (Games > exams atm). Much, much later (about an hour ago, based on when I’m writing this ), I went to see what the story was. Load up FF3 and, bam, 500 error. Load up Windows Live Mail, bam, no support response waiting in my inbox. Now we’re starting to get a little cheesed off…

So pretty much the last hour has been spent Googling around to try and find answers to this problem, or at least to read the tales of other people who have had the same problem as me. Most of the results were from people using the WP-Cache plugin on their blogs which seems to clash with my host. However there was one result in which someone said that there may be a problem with the .htaccess file causing problems. I decided to take a chance as see whether there was anything in my .htaccess causing this 500 error.

Mod_rewrite stuff – Nope, not causing it.
AddHandler commands – Negative.
Huge long list of deny lines – No…. hang on a mo, there’s the culprit.

So, it’s making me think how on earth that huge long list of IP addresses that are banned from accessing my site is causing the problem. I remember that earlier today I had to ban a new IP, and I had added a comment to explain why it was banned. So, to give it a shot, I put the IP’s back and removed the comment. Success!

But why the comment was causing the error was beyond me. It was commented directly after the deny xx.xx.xx.xx, so there shouldn’t have been a problem. I put the first comment back and it worked. I put the second back and it broke. At least now I had managed to find the exact culprit. And then came the most annoying bit out of this entire fiasco. I replaced a backslash I had in the comment with a space…

…and the fucking thing worked again…

For fuck’s sake… all this downtime for the sake of a god-damn backslash…

Oh BTW, I still haven’t received a response from the support team…