Saturday January 1, 2000 - 9:49pm
It appears that we successfully rolled over to 2000 without a single glitch!
A major DNS server change occurred yesterday and was finished up today. There
was a period of downtime for about 5 of our clients caused by a very simple error on our
part (forgetting a semi-colon). I also forgot to note that about one week ago we
added secure server support that will soon be available to all clients, free of charge.
Wednesday November 24, 1999 - 3:02am
I became rather bored, so I posted some pictures of our server equipment here. Yes, our server room is very messy and not much to
look at, yet...
Saturday November 20, 1999 - about 2pm
We had to take one of our web servers down for unscheduled maintenance this
afternoon to replace a faulty motherboard. The motherboard was causing the machine
to lock up on occasions. We also updated our web server software to add database and
scripting functionality. This functionality is needed for the web-based control
panel we are still working on in-house. The server, which hosts a majority of our
websites, was down for about 10 minutes on 2 separate occasions. (Tech
details: RedHat's stock Apache 1.3.6 installation was replaced with a custom compiled
Apache 1.3.9 with PHP3. MySQL was also added for database support.)
Tuesday November 7, 1999 - 10:03pm
Over the next few weeks, we will be installing 3 security software packages on
our UNIX systems. These packages are designed to take an "active" role in
system security. The system will disable accounts that it thinks are being attacked.
We will fully test this before putting it on our production machine. If you
experience problems with logging in and suspect it may be because the security system
locked your account, you can contact firstname.lastname@example.org
and we will resolve the issue immediately.
Tuesday November 7, 1999 - about 1:40pm
We have been working to maximize our security on our servers over the past few
weeks. Today, our system administrator made a minor mistake on a firewalling rule
that prevented any access to our server from the internet. Our UNIX webserver was
not responding to requests for about a 11 minute period while we traced the problem and
Thursday October 21, 1999 - 10:05pm
This evening we have replaced the 8-port hub connecting our servers with a
24-port switch. This upgrade was purely to improve performance on the network.
During the change over we experienced 2 or 3 outages, two very short ones (~40 seconds)
and one longer one (about 5 minutes.) Our apologies for not providing notice on this
Thursday October 14, 1999 - 1:27pm
Our mail server was down last night around 1am eastern for about 5 minutes to make some
necessary changes for the web-based control panel for users. This may occur again
tonight if necessary. Please note that e-mail will not be lost during this period.
Monday October 11, 1999 - 3:01pm
The http service on aurora, our main UNIX webserver, caused some interesting problems
around 1:50pm today. One of the processes became "zombie" and somehow made
our access and error logs enormous in size, taking the system's partition from about 40%
usage to 100% almost immediately. The problem did not cause an interruption of
services, just a slow down. The web services were down for about 2 minutes while we
removed the logs to free up disk space. We were paged immediately when the disk
space usage reached alert level.
Sunday October 3, 1999 - 1:03pm
We just changed to a new FTP server software on aurora (a.k.a. ftp2), the primary UNIX web
server. This change is mainly for some security concerns. Users should notice
little difference. The change brought FTP services down for about 4 minutes.
(Technical info: The change was from wu-ftpd to ProFTPd)