Wed Apr 12 20:29:51 PDT 2000 — The new circuit to Cable and Wireless is UP, and the network has converged. This should improve our connectivity greatly, and we’ll notice even more performance benefits as we tune the circuit and routing tables over the next few days. We apologize for the transition hiccup, as this was designed to be a seamless operation. We’ve all been working hard to get this circuit turned up ASAP, and after weeks of work to do so, there is a fair amount of jubilance among our staff! If you have any questions, please post to news:sonic.net — Eli (The excited guy who got to actually turn up the circuit in the Cisco), and the entire Operations Staff at Sonic.

Wed Apr 12 20:03:10 PDT 2000 — We are working with Cable and Wireless to get our new T3 installed and running, this is causing some spotty performance on our other T3. Please bare with us on this and this will clear up shortly. Thanks – Eli and Steve

Tue Apr 11 11:50:08 PDT 2000 — We have implemented a newsfeed filter to help clean annoying SPAM, excessive cross-posts, and other garbage from the news server. In just the first few hours its already blocked out nearly 3000 bad posts. If you are interested in more information, I’ll be posting to sonic.general -Kelsey

Mon Apr 10 14:05:20 PDT 2000 — The replacement cache adapter arrived this morning and was promptly installed. However, it has been exhibiting the same behavior as the old hardware. After many calls to Cidera we appear to have stabilized the cache adapter and are awaiting a second replacement. Meanwhile, we added 2GB more spool for the *jobs* groups and 2GB more to sonoma.* and sonic.* to increase the article retention time in those groups. It appears that text groups are keeping articles for around 10 days and binary groups for around 3 days. We are continuing to work on increasing our article retention times and on developing some news SPAM filtering. -Kelsey and Eli

Fri Apr 7 12:52:14 PDT 2000 — We have taken the news server offline to initialize the new spool space that we added a few days ago. It is going to take about a half and hour to ‘create’ the spool files. We did not anticipate that it would be offline for more than just a few minutes or we would have done it in the early morning. Once it is back online it will have an additional 30GB of text spools and 90GB of binaries spool. We’ve also taken a few steps to optimize performance. -Kelsey and Eli

Fri Apr 7 18:41:49 PDT 2000 — The ‘cache adapter’ appears to be failing. It is black box hardware so there isn’t anything we can do until the the replacement arrives. Unfortunately since it is Friday night we probably won’t receive a replacement until Tuesday with a slight chance that it could arrive on Monday. Our terrestrial feeds are working fine but are only able to bring in about half the volume that the satellite link can. It’s ironic that this happened today after Eli and I had rushed to add 120GB of spool to the news server for the weekend. -Kelsey

Fri Apr 7 15:40:36 PDT 2000 — Our cidera satellite news feed was offline and we had to reboot the ‘cache adapter’ server to get the feed back online. As a result, our news volume may be a little less than normal today. Hopefully our five terrestrial feeds will be able to fill in all the missing holes. -Kelsey

Thu Apr 6 10:00:46 PDT 2000 — The current multihome web server, thunder.sonic.net had problems this morning at around 5am. Scott and I both came down to reset it, and all’s well. As it turns out, Lyons really is NOT a 24 hour deal – we had to go to Dennys afterwards for breakfast. =) Kelsey, Scott and Eli have been working hard to get our load balancing stuff all stable, and we seem to be there on the SMTP and POP now. I’m hoping that we can wrap up the web server load balancing in the next day or few, and once this is complete, a failure like this won’t be noticed at all by our customers. -Dane and Scott

Thu Apr 6 17:46:54 PDT 2000 — We are noticing a number of disturbances out on the backbone, mostly centered around CRL at this time. This problem may affect Sonic customers trying to access remote sites. For info, check out mdr.ihr.daze.net/ – Also note that while this affects our connectivity, it also affects the rest of the Internet and we don’t have any control over when it gets repaired. — Eli, Kelsey

Wed Apr 5 17:39:55 PDT 2000 — We’ve been getting more reports of two mail problems: 1) Mail being bounced with ‘user unknown’, and 2) mail to domainaliases being delivered to ‘real’ sonic accounts. To address this issue we analyzed logs on all four mail servers, and determined that (so far, knock on wood) Prop hasn’t had these problems. We’ve just completed duplicating Prop’s configuration on the other three mail servers: Sub, Ultra, and Buzz. Hopefully this will alleviate the problems we’ve seen. -Scott and Kelsey