Month: March 2007

Emergency SQL Maintenance.

Mon Mar 26 23:52:40 PDT 2007 — Emergency SQL Maintenance. Our customer MySQL and Postgres database server is undergoing emergency maintenance to replace a failing disk and suspect controller in it’s local RAID. Due to some complications we may incur a few minutes of downtime as the hardware is swapped out. -Kelsey and Nathan

Update Tue Mar 27 01:22:57 PDT 2007 — The customer SQL server emergency maintenance went off without a hitch. Actual down time was less than 8 minutes in total. -Kelsey and Nathan

Dial move.

Tue Mar 20 11:01:25 PDT 2007 — Dial move. At approximately 10:00am on March 23rd we will be migrating a piece of our dial-up modem capacity to a new provider. While we expect little to no customer impact from this swap, it is possible that the moved numbers will briefly fail to function. No in-process calls will be dropped, and we’ll have increased technical support staff on-hand to handle any problems as they come up.

All dial-up numbers ending in a “3” will be affected by this migration. Redundant, unaffected numbers are available in all areas. To find them, browse to:

-Nathan and Eli

Santa Rosa colocation stays up.

Mon Mar 19 12:44:54 PDT 2007 — Santa Rosa colocation stays up. Early Friday morning we experienced an outage on one of the two AT&T-provided Gigabit links leaving our Santa Rosa datacenter. Our Network Operations Center worked to isolate the trouble, which turned out to be a failed regenerator card in a Fremont central office. The card was replaced around 4:00pm, bringing the link back into service. The end result was a total of 2 seconds of packet loss for colocated customers, plus a very happy NOC staff. It’s always fun to see our carefully planned redundancy work in the real world! -Nathan and the NOC (Jared, Zeke, and Matt)

PG&E Service Outage at our Santa Rosa…

Thu Mar 15 13:34:37 PDT 2007 — PG&E Service Outage at our Santa Rosa datacenter and colo facility. We lost PG&E electrical services briefly this morning; Our backup power systems functioned perfectly. By the time we’d made it to the electrical room our genset had already been automatically started and the ATS had just switched over to it. There was no interruption of any services in or provided by our datacenter. It was exciting and reassuring to see everything work exactly like it is supposed to! Operations Staff

Partial DSL outage.

Wed Mar 14 19:04:41 PDT 2007 — Partial DSL outage. A software problem in our provisioning tools caused a ‘sync no surf’ condition for about 8% of DSL subscribers. The duration was about 15 minutes. We are working to identify and correct the bug as we speak. – Eli, Nathan, Ross

Sacramento Area DSL Circuit Upgrade.

Mon Mar 12 12:47:39 PDT 2007 — Sacramento Area DSL Circuit Upgrade. Early Wednesday morning at approximately 12:01 AM, the circuit that handles all DSL traffic for the Sacramento area is going to be migrated to a new, larger circuit. Customers may experience up to 30 minutes downtime as they are migrated in batches from the old to new circuit. -Nathan and Kelsey 😉

MOTD Correction.

Thu Mar 8 19:56:02 PST 2007 — MOTD Correction. The DSL maintenance the MOTD reported as being scheduled for early Saturday Morning is actually going to be taking place in several hours, not tomorrow night. We apologize for any confusion. -Kelsey

Planned power and DSL maintenance.

Wed Mar 7 16:23:26 PST 2007 — Planned power and DSL maintenance. During the day tomorrow we will be doing DC power plant maintenance at one of our colo facilities in San Francisco. There should be no associated services interruptions of any kind; however, there is the potential that power could be lost to some of our DSL termination equipment. In the event that power is lost to any of our equipment the associated outage would be brief.

Early Saturday at 12:01AM, a number of planned events may affect a number of short outages to some DSL customers terminated in San Francisco. Jobs to be done include, installation of additional ATM switch ports, turn up of additional DSL termination equipment, and restoration of one of the ATM OC-12s to it’s primary path through our in building patching and switching. Affected customers should see a total of no more than 15 minutes of combined outage spread over a number of short events. Network Operations