IsMyWebsite Free Hosting Development Blog

It’s been a long time since I posted the last update, but there hasn’t really been a lot of problems. So overall, excellent to see.

Node 2 – Felweb
Uptime last 14 days – 100.00% (Up 0.31%)

Node 4 – AquariusStorage
Uptime last 14 days – 100.00% (New Node)

Node 7 – HighLayer
Uptime last 14 days – 99.84% (Down 0.16%)

Node 6 – Nixism
Uptime last 14 days – 99.84% (Down 0.16%)

Node 5 – Addora
Uptime last 14 days – 99.84% (Up 0.16%)

Node 3 – OSHS
Uptime last 14 days – 99.69% (Up 5.92%)

Node 1 – SmokyHosts
Uptime last 14 days – 98.43% (Up 0.51%)

Please let us know if you ever need any assistance with your hosting.

Some people still don’t get what happened. Here’s an analogy:

Imagine if every Monday, you ask how to get to school. But one day the information booth is closed. So you give up and collapse at the side of the street in a heap. That’s what any computers that couldn’t access our domain or your websites did.

And also consider this. That whole thing happened because the person operating that information booth didn’t have directions to one location. They didn’t know how to get to the theatre, let’s say. That operator doesn’t know how to get to the theatre, so he refuses to come to work, and everyone collapses in heaps at the side of the road because they don’t know how to get where they need to go. Not only that. He doesn’t call in sick. He doesn’t complain. There aren’t any errors. He just doesn’t come.

Sound absurd? Well, that’s why I’m a little bit more than upset about what happened. I checked. I made sure school was open for the day. I made sure the teachers were ready to teach. I made sure that the grass was cut, the flowers watered, the gates open. But none of that mattered. The guy at the information booth wouldn’t tell anyone how to get to the school, so they didn’t go, even though it’s at the same place it’s always been.

Well, guess what? I’ve hired someone new to man the information booth, and this guy WILL tell me if things go wrong and he can’t come to work, something that has never happened since we first created the DNS system 9 months ago. I’m also building a second information booth, so there will be no collapsing in a heap by the side of the road for anyone anymore.

Some people are leaving, and I guess that’s fine. They have a right to be upset. But not my fault everyone decided to collapse in heaps at the side of the road. I only forgot to tell the information booth operator where the theater was, one line in over a million lines of DNS configuration. And it wasn’t even that I didn’t specify that. It was a glitch in the commercial control panel I used to set the DNS. The entry disappeared after I set it.

That’s DNS for you, one of the most frustrating and difficult to manage systems on the face of the planet. There are probably only a few thousand people in the entire world who even know enough to build a system like ours, so I’m sorry that it took a little over a day to fix it.

Any complex system that accepts dynamic input is subject to the possibility of wrong output. Any precautions would be to prevent the occurance, and improve the resolution.

Prevention:

• Once this situation is resolved, build the solution into the system, so that any future cases are not subject to human error.
• Not use the commercial control panel for DNS modifications.

Resolution:

• Create an alert to be generated anytime to DNS regeneration has resulted in an error. It would mail me automatically with all applicable details. I would immediately know what and where the problem is.
• Keep a log of the DNS launch sequence, which can be loaded among detection of the error, and contains full details.
• Store DNS entries in alphanumeric order instead of by ID number. This will make it faster to find the problem zone among the thousands of entries that are currently active.

In 9 days, we mark the first anniversary of the most successful system implemented throughout our 3 year history. For many of our clients, this will be the first time they have ever experienced something like this.

On the plus side, I did get to meet quite a few of our clients today for the first time. I’d really like to thank everyone who took the time to let me know their little tidbit of information. That was all quite helpful to put together the full picture of what was going on, and I hope everyone will continue to provide feedback all the time, and not just when things go horribly wrong.

I just want to thank everyone who’s still here. Interestingly enough, our ratings have been HIGHER as a result of this situation. So thanks everyone for all your wonderful support and patience as we worked to resolve things.

For those curious about the details, I can give you a few of those. When you visit ismywebsite.com, or something.ismywebsite.com, or anything.ismywebsite.com, your connection is first routed to our DNS server, which basically tells you where to find the resource. Then, you look at that server.

Most of the time, your computer only needs to look this information up once. It just remembers from the last time. It caches the data and goes to the same place time after time. But, as you can imagine, this creates a problem when a website moves to a new server, so that’s why your DNS records are updated every 24-48 hours.

As a DNS server does not need to be visited every time you visit the site, downtime and unavailability SHOULD have no effect if you’ve already visited the site. The real mystery is, why our computers all eventually declared mission failure without even checking the CACHED location, which was up and completely operational throughout this time. Yet this is apparently how DNS is implemented in most networks.

So, sometime early yesterday morning I was messing with the DNS for one of our clients who wants to use their IsMyWebsite.com domain on another hosting provider, but the other provider isn’t cooperating very well. After some degree of fidgeting, the DNS ended up with an unusable configuration, but no errors or indication of any kind was given by the commercial control panel we are using.

Yesterday was a very busy day. I had 2 assigments, a quiz, and a midterm all within the one day. So, when some support mentioning our server being down started to come in, I did a quick check and found only the main domain (ismywebsite.com) down, and all servers online, so I assumed it would sort itself out since the main site server is managed by VectorLevel to maintain a high degree of uptime at all times.

Finally, last night I spent a total of 2 hours playing around with the configurations, before noticing that the DNS server’s main domain was down. Then I spent 3 hours struggling to figure out what was wrong with that server. Finally, at 5 AM, I resigned myself to the fact that a half night’s sleep would best enable me to function for the rest of the day.

And I was right. When I woke this morning, I managed to solve it within 30 minutes, thus ending 36 hours without DNS functionality, the first time anything of this nature has happened in recent history.

I’ve just spent the whole morning practically answering the endless bombardment of emails. I haven’t even been able to catch up on the latest website requests to come in… This is basically every day, but none of you can really see what I do because I only help each of you for 5 minutes or so.

I’m sorry my donation requesting email wasn’t perfect and sounded like some kind of sales pitch or begging. Thanks to those who asked to have their accounts deleted and all communication with me removed for good, and double thanks to those who came back with comments such as the one I got here:

make money for us instead of relying on donations.

Make money for you? Is that like the thousands of dollars of my own personal income that I’ve poured into this hosting over the past 3 years, and am probably never going to get back, or is it more like getting meaner with the ads policy and dealing with comments like these:

“i hate adwertising. !!!HATE!!! !H!A!T!E!”

“Well, the ads is in any case a limitation”

Anyways guys, I’m sorry I tried. Really. If you can’t afford to help out with financing, that’s fine. If you can’t possibly place ads on your site, whatever. If you can’t find the time to stay active on the forums, I understand.

But surely you have to be able to help with something…

Otherwise this just isn’t going to work…

Lots of changes in uptime so quickly. Here’s the latest:

Node 7 – HighLayer
Uptime last 14 days – 100.00% (Up 0.23%)

No current issues. Not open to new accounts.

Node 6 – Nixism
Uptime last 11 days – 100.00% (Up 3.18%)

No current issues.

Node 5 – Addora
Uptime last 14 days – 99.69% (No Change)

cPanel license error resolved.

Node 2 – Felweb
Uptime last 14 days – 99.69% (Up 0.23%)

Uptime tracking now back online.

Node 1 – SmokyHosts
Uptime last 14 days – 97.91% (Down 0.75%)

No current issues.

Node 3 – OSHS
Uptime last 14 days – 93.77% (Down 6.23%)

We are planning to discontinue this node in the near future.

Please let us know if you ever need any assistance with your hosting.

Lots of changes in uptime so quickly. Here’s the latest:

Node 3 – OSHS
Uptime last 14 days – 100.00% (No Change)

Storage will not be upgraded. Available to donors only.

Node 7 – HighLayer
Uptime last 14 days – 99.77% (Up 2.04%)

No current issues. Not open to new accounts.

Node 5 – Addora
Uptime last 14 days – 99.69% (Down 0.31%)

cPanel license error resolved.

Node 2 – Felweb
Uptime last 14 days – 99.46% (Down 0.54%)

Uptime tracking now back online.

Node 1 – SmokyHosts
Uptime last 14 days – 98.66% (Up 0.58%)

No current issues.

Node 6 – Nixism
Uptime last 14 days – 96.82% (Down 0.51%)

No current issues.

I’ve just been talking with the administration of Node 6, and it appears no move has actually been completed yet at this time. Conditions are expected to improve once they are migrated to the new datacenter, however no date has yet been set at this time.

Please let us know if you ever need any assistance with your hosting.

I’ll be getting some download speed tests underway here soon, but for now just another uptime update.

Node 7 – HighLayer
Uptime last 14 days – 100.00% (No Change)

No current issues. Not open to new accounts.

Node 3 – OSHS
Uptime last 14 days – 99.84% (Down 0.16%)

Storage will not be upgraded. Available to donors only.

Node 5 – Addora
Uptime last 14 days – 99.69% (No Change)

cPanel license error resolved.

Node 2 – Felweb
Uptime last 13 days – 99.66% (Up 0.39%)

Uptime tracking now back online.

Node 6 – Nixism
Uptime last 14 days – 98.41% (Up 0.15%)

No current issues.

Node 1 – SmokyHosts
Uptime last 14 days – 98.16% (Down 0.15%)

No current issues.

Please let us know if you need any assistance choosing the best server for your hosting.