Friday, March 23, 2012

Email Issues

MailScanner released an upgrade today (v4.84.5), which may have a bug in it (there have been 2 versions released today) and MailScanner is refusing to start/stay running.

While we debug this issue MailScanner has been disabled. Email will still be delivered, but you may see a slight delay in delivery as we stop and restart email processing services in order to work on this issue. You may also see spam slipping through until MailScanner is fully functional once again, as that's MailScanners' sole purpose in life: to filter spam.

We're working on fixing this issue as quickly as possible, and will update this blog post when the issue is resolved. Thank you for your patience.

EDIT 8:18pm: The MailScanner issues have been resolved and it is once again humming right along, like a finely-tuned spam-eating machine.

Wednesday, March 14, 2012

Brief downtime to repair server quotas

The quota system on the new Apollo server has not been reporting correctly. It was supposed to be fixed when the server was first commissioned, but I noticed today that it's still not correct.

The data center is running a fix on it, which will incur 2-10 minutes of downtime. We apologize for any inconvenience this may cause.

Monday, March 12, 2012

Post-Migration Issues Being Cleared Up

Like true love, the course of server migration doesn't always go as smoothly as planned. While most users didn't notice even a hiccup, we're currently working on repairing a few permission and database issues that happened when transferring accounts from the old Apollo server to the new Apollo server.

1. If you're currently unable to reach your website (or ours), it is likely a DNS caching issue, which can happen when the server IP address changes. Just like when a browser will cache a copy of a web page that may have recently changed, browsers also store a DNS cache (depending on your settings). To flush your local computer's DNS:

On Windows:
  1. Open the Start menu.
  2. Go to Run.
  3. If you do not see the Run command in Vista, search for "run" in the Search bar.
  4. In the Run text box, type: ipconfig /flushdns
  5. Press Enter or Return, and your cache will be flushed.

On Mac:
  1. Go to Applications.
  2. Go to Utilities.
  3. Open the Terminal application.
  4. Type: dscacheutil -flushcache
  5. Press Enter or Return, and your cache will be flushed.

After flushing your local DNS cache, it's also a good idea to reboot your computer.

I personally use and recommend using Google DNS instead of your ISP's DNS. I've found their DNS servers to be very fast and super responsive to DNS changes. Here is one tutorial on how to make changes to the DNS server your computer uses.

A caveat: Nothing is foolproof. This morning I was able to view our own status blog (and a post on it made this morning, so it wasn't a cached page) on one computer that doesn't use Google DNS, while from another computer on the same network/IP -set up to use Google DNS- could NOT see the status blog. My computer at home, also using Google DNS but on a different ISP/network/IP couldn't see our status blog.

These issues were cleared up when I resynced the blog feed (through Feedburner), which forces a recheck of the current DNS settings. The funniest part is, Feedburner is owned by Google. :)

2. Check to see if your site is up
Here's another quick way you can check to see if it might be a local caching issue. There are many websites that will check to see if your website is up, here's a couple:

3. Use a website monitoring service
Another recommendation is to use a website monitoring service, such as Pingdom or Binary Canary that alerts you if your website is down. Both offer free accounts if you only need to monitor one website. This would be another tool to help you know if/when there's a problem, and if it's one with your network/ISP/local computer verses a problem with your website or our server.

4. If you're receiving an error message on your site or when logging in to cPanel it speeds things up considerably if you can give us the error message itself. There's a million things it could be, and a hint as to what error you're getting will help considerably.

5. Apache is currently being recompiled has been recompiled to include the PDO MySQL extension. That was somehow missed when the settings were transferred.

6. If you're experiencing any other issues we're currently available by Live Chat (the icon in the bottom right corner of all pages on our website...if you don't see the Live Chat icon, refresh the page), or you can submit a ticket through our online support desk.

If you ever can't access the support desk/Client Portal for any reason, you can also send a Direct Message to our Twitter account and that will send a smoke signal. Per how Twitter works, you need to be following us and visa versa.

Friday, March 9, 2012

Server Migration In Process

Services such as email, cPanel and FTP will be shut off for the next 30-60 minutes or so while we migrate accounts to the new server and transfer the DNS settings. We don't want you to lose any email in the process of the transfer, so that's why we do it this way.

This post will be updated when the process is complete.

3/10/12 1:41AM
Well, that was a long 30 minutes, wasn't it? There were a few discrepancies between how cPanel is supposed to transfer user accounts, and how they actually did in this case. We were only able to transfer the accounts one at a time, due to a bug in cPanels' "transfer multiple accounts" option.

DNS has been transferred, the email queue on the old server was flushed (all mail holding during the transfer was released to it's intended recipients), and our backup DNS provider is likely earning their pay about now.

We'll be babysitting the process over the next few hours and tomorrow, to make sure everything runs smoothly.

The server SSL certificate will be reinstalled on the new server as soon as DNS is done (it's assigned to the hostname, not the IP), so if you try to access email and get a certificate warning, it's because the server is using the default self-signed one that comes with cPanel, instead of our spiffy official one.

Please be patient as we make the final adjustments and tweaks.

Thursday, March 1, 2012

Email Issues Solved

Apparently ClamAV, the email virus scanner, didn't like leap year. At approximately midnight the mail server stopped processing email because the virus scanner stopped. This issue was corrected and all email holding in the server queue was released at 8:00am.

We apologize for any inconvenience this may have caused and will investigated further to make sure this doesn't happen again.