GuaranteedVPS - 1 (888) 466-2132

On Friday 30 November 2007, an unexpected carryover from yesterday’s rack transition incident caused a major outage for all of our servers. A neighboring rack blew a power coupling while two of the three redundant fiber switches were being relocated. This damage to the power coupling also damaged the remaining fiber switch. The switch was patched as quickly as possible. Ironically, it was the case that the relocation of the fiber switches was being done in an effort to put them on their own coupling to eliminate any chance of what actually happened.

While our servers actually remained on and up, there was no route to your hosts between 8:36AM and 10:34AM PST. The situation was resolved at 10:34AM PST today, and a more permanent solution will be in place by 12:01AM tomorrow, Saturday 01 December. Due to this permanent solution, there will be another downtime which is expected to last less than 5 minutes, and no more than 15 on the outside, at approximately just before midnight tonight.

You may have noticed that your servers have, in fact, reported a downtime. We used the debacles of late to go ahead and migrate the servers to their new and better homes since there was no route to your hosts from the public internet anyway. This allowed us to be more efficient within the parameters of the outage and prevented us from having to reschedule the downtime to relocate the machines later.

This issue is an isolated incident, and was the result of our demand of our upstream provider to make sure that an incident like this never actually happens. If you have any questions or wish to discuss further any of the issues, please do not hesitate to open a support ticket by either calling us directly or via email.

AngryBox will be undergoing a scheduled downtime Monday night to add a few modules for a customer. Nothing is wrong, but adding the modules requires an environment reboot. This notice does not affect any other box. Expected downtime is less than two minutes.

AngryBox and CurmudgeonBox have been updated and are running without a hitch, though it took about five minutes longer than expected. Kernels are up to date.

AngryBox will be going down briefly in eight days, in order to perform kernel and security updates and other upkeep tasks. Expected downtime is on the order of three to five minutes. We will be doing this at an hour so late that some countries don’t even have it (better known as two in the morning, Pacific Daylight Time.) This is standard housecleaning, and happens for an average server twice a year; nothing is wrong. AngryBox is, in fact, chugging along nicely. We’d just like to make sure it’s up to date.

Please make sure that any server software you’re running is set to start on boot automatically. If you are using a standard image, which almost all of you are, then Apache and SSH are already set up to boot; some servers, such as Exim, MySQL and Postfix, automatically set this up, whereas others, such as Qmail, PostgreSQL and Erlang, do not. Almost all services will be enabled/disabled through the unix command “chkconfig.” If you need help getting ready for reboot, let us know, and we’ll toss in a hand.

GuaranteedVPS: We Patch Stuff. ™

It’s hard to come up with a funny byline for routine maintenance. Sorry.

There was a four minute outage of network access to AngryBox (most .79.* customers) while a switch at the data center was replaced. We apologize for the inconvenience, though it doesn’t seem that anybody actually noticed. There was no loss of data and nothing is wrong with the machine; just that one of the upstream switches went kaput.