Month: January 2005

Tue Jan 11 08:07:21 PST 2005 — Broadlink Wally Moore tower offline. Broadlink reports their Wally Moore site is currently offline. They suspect power problems. If that is the case, they estimate an ETR of 30 minutes. -Scott, and Tim from Broadlink

Tue Jan 11 14:22:04 PST 2005 — Broadlink TMAC and Airport towers experiencing latency. Customers in the northern part of Santa Rosa are seeing latent connections and ~25% packet loss. Broadlink is diagnosing the issue, we will update the MOTD when we have an ETR. -Eli, Jason Kane Update – 16:29 PST – Issue Resolved. Jason traced the problem to a cable that had taken on moisture, and the replacement cable has restored service. – Eli, Jason

Fri Jan 7 17:16:45 PST 2005 — Kernel Security Updates. All public access servers have been patched against the recently announced local root exploit in the Linux kernel. -Kelsey, Nathan and Scott

Fri Jan 7 15:28:21 PST 2005 — Datacenter power work Tuesday, 1/11/05. PG&E will be replacing the transformer for our building at 8 AM Tuesday. We are renting a generator to provide power to the building during the work period. Customers should see no change in service during this work. The datacenter itself is on UPS, and if the rental generator fails, our own standby generator will take over. Our tech support systems have their own UPS’s, so those friendly folks can always be there to help. -Russ and John

Fri Jan 7 15:22:35 PST 2005 — Colocation UPS enhanced. Today we added a second, parallel battery string to the UPS that serves our Santa Rosa datacenter. This doubles our backup run time, and eliminates a single source of failure that caused problems in the past. -Russ and John

Mon Jan 3 11:32:52 PST 2005 — News Feed Issues. Early yesterday afternoon our internal news feeder server started to reject all posts due to a file system problem. This was not picked up by our monitoring. When we discovered it through some user reports early this morning we restored services for a couple of hours until it failed again. All of the problems were resolved at 10:50 AM. This failure will result in gaps on news.sonic.net as well as the potential loss of outbound posts. This only affected news.sonic.net and not supernews.sonic.net or newscene.sonic.net which were both available and fully functional. -Kelsey

Sun Jan 2 22:20:28 PST 2005 — Sluggish performance on some dialup connections. Some customers reported poor connectivity when dialing into one of our POPs. Rebooting the access servers has corrected the problem. The affected POP serves Sonic.net access numbers ending with 3, other access numbers were not affected. -Nathan and Russ