Yesterday am a Verizon tech working on the place next door yanked out all of our circuits. This put our datacenter dark and took us off the Internet across all of our lines. Because it took out our DNS servers even the SSH servers and such were unreachable.
We’ve had no luck getting Verizon back out here today. They move at their own pace and we’re two services back from them. We shouted at anyone who would listen, tried to escalate, tried to get a weekend visit once 5pm hit, but no luck on any. It looks like we will remain down until Monday…on their techs error…I am furious.
We have since added additional dns servers and are working at getting the SSH servers to again respond to logins. Currently the update has propagated and the web hosting servers, irc, and socksplus are now available. Soon the rest of the SSH servers will be as well. Webmail and pop/imap/smtp look like they will remain down until Monday.
I’ve completed the new fetch mail feature, it is in limited beta now. If you’d like to try it out send me an e-mail. Basically it is a way to get e-mail from any other mail provider (even yahoo without a paid pop account).
Most of the popular e-mail services are listed in a self configuring droplist. You also have the ability to add custom servers. It can connect on any port, with any popular authentication method, any of the common protocols (pop, pops, imap, imaps, httpmail), and it will automatically use SSL/TLS if the server supports it.
Mail can be scheduled to be retrieved from other accounts both on demand and even when you are not logged in. So set up your other accounts and then you’ll only have to connect to cotse via pop/imap/or webmail and it will all be there, your other mail fetched automatically throughout the day for you.
All fetched mail passes through our filters. This enables you to apply all our filters to your other e-mail accounts. Headers are added so it’s easy to key on which account fetched what in any custom filters or redirects you’d like. It will also provide you with detailed reports of the transaction if you wish.
It is infinitely configurable and has more features for connecting to different e-mail sites than most mail clients. The other plus is that all those sites see is a cotse IP fetching mail, so it shields you from them as well.
I’ll run it limited beta for a while and hope to move it live to a site wide beta within a week or so. Again, if you’d like to try it now, send me an e-mail.
I am writing less frequently because I am coding. I’m working on adding the ability to retrieve mail from nearly anywhere. It will retrieve POP3/POP3s, IMAP/IMAPs, and HTTPMail. Retrieval can be set to be automated, regularly retrieving even when not logged into webmail. Retrieved mail will be passed through existing filters.
This means users will be able to set up accounts to regularly check. Such as Hotmail, Yahoo, Juno, WebTV, Netscape, ISP mail accounts, other hosting accounts, and and any other account someone might have. Settings can be made to automatically retrieve it every 30 minutes, every hour, every 6 hours, or once daily. All cotse’s filters will work on the fetched mail, so someone can spam filter it and direct it to a folder or other mail address as with all cotse e-mail.
The end result will be all mail in one place. Then all someone will have to do is check their cotse mail either via webmail or their favorite pop or imap client and they are getting mail from all their accounts. Even pocketPC, Palm, and other portable devices have pop/imap clients. Set it and forget it.
The interface is about 70% and so is the back end. Then I have to tie them together. But it will be a reality to at least make beta sometime soon. Depending upon how much coding time I can get.
4 am, the sound I least want to hear, my phone paging me. It was the monitoring software telling me there was a problem with the mail gateway. Straight out of REM sleep so deep I couldn’t see for a little while I move to my computer to check on it.
Apparently the gateway hung and as should happen it was powercycled by the monitoring software when it could not connect nor respond to ping. It came back and responded, but all connections to it just hung indefinitely. It accepted them fine, but they all just hung. Even the remote console was sluggish, making troubleshooting more difficult.
Turned out to be corrupted data that hosed NFS, a manual fsck (the auto fsck on reboot didn’t do the trick apparently) boot back into multiuser and all seems to be well. If you could not connect for a brief period, this was why. Back to sleep.
We’ve had to get strict on credit card acceptance. This is because of the amount of fraud on the net. Nobody gets hurt by this fraud but the merchant. This is because the bank takes the money back and then charges the merchant an additional $25 for doing so. The consumer isn’t responsible for the charge, the bank isn’t hurt, the bank actually makes money on fraud with that $25 fee, only the merchant is hurt, they lose not only the money but an additional $25.
Because of this we can’t just accept whatever cards the bank approves. If we did that we’d go bankrupt in chargeback fees. So we developed our own process. We have a number of ways we verify charges before submitting them. One of the most effective has been placing required information into a page and making that page required reading. If the page isn’t read, the person doesn’t know that the information is required and omits it. Most of the carders don’t read it, so it’s a quick first level weed out when the required information detailed on that page is missing.
But it also catches some legit. This I can’t quite comprehend. This means that there are people who think enough about privacy to sign up with a privacy service, yet not enough about their privacy to read the page where they are entering their credit card details. They are entering their personal information into some page they didn’t even read, yet they “care” about their privacy? This makes no sense to me.
