This morning, at 8:32 AM, one of our customer MySQL servers crashed due to an apparent bug in MySQL. Additionally, after the crash it was failing to restart. The trouble was isolated to a specific user’s database, which when manually removed, allowed the MySQL server to start normally. The server appears to be functioning correctly at this time.
We have confirmed with AT&T that there is currently a DSL outage in the San Jose area. They are working to resolve the issue but we do not currently have an ETR. Symptoms are that customers have sync but are unable to get on line.
Update: The outage should be clear. If your connection is still not working please try rebooting your DSL modem.
Tonight, Tuesday November 29 at 12:01 AM, we will be performing a maintenance reload on several legacy DSL aggregation routers that serve some of our customers in the Bay Area and Los Angeles areas. Affected customers will experience 5-10 minutes of downtime as the routers reboot.
Update: The maintenance reloads have been completed and all affected customers should be back online.
This evening, our support team began tracking an issue causing poor DSL performance for legacy DSL subscribers served by a certain SMS. Upon further investigation, we found a port on our ATM switch to be failing, and causing loss for traffic going across it. We moved the SMS to a different port on the ATM switch, and the performance problem should be resolved now.
-Jared and John
Earlier this afternoon, we made some changes to the authentication service used by our VPN Concentrator, which resulted in an inability to establish new connections to the VPN Concentrator. We tracked down and fixed the configuration problem that caused this issue, and all VPN sessions should be able to connect at this time. We apologize for any inconvenience this caused.
-Jared and Jasper
At approximately 3:09PM, while performing routine cable maintenance, a telco technician accidentally pulled an incorrect fiber jumper causing a brief outage. This would have affected legacy DSL service for customers in the Sacramento area. The mistake was immediately corrected and service was restored less than a minute later. We apologize for any inconvenience this may have caused.
-Tim and Nathan
The raised floor bracing and cabinet mounting systems at our Santa Rosa datacenter are currently undergoing a seismic retrofit. The work is being handled by a contractor that specializes in raised floor environments and will not cause an interruption in any of our services. Colo customers should note that some cabinet doors must be opened in order to provide sufficient clearance to perform the work. Please do not be alarmed at an unexpected door open event during regular business hours over the next few days.
Tonight, Friday November 4 at 12:01 AM, we will be reconfiguring approximately 50 Flexlink Ethernet customers to smooth the way for future network upgrades. This work will disrupt connectivity to those customers for less than 5 minutes apiece. No work or changes will be necessary for these customers to restore service.
-Jared, Tomoc, Robbie, and Phil
Update: All customer reconfigurations are complete. No customers should have been offline for more than 5 minutes. If you are experiencing any problems with your Flexlink Ethernet service, please call the NOC Hotline at 877-706-5662.
We were notified this morning of issues dialing numbers in Northern California on our Fusion Voice service. The issue was traced to one of our upstream phone providers in this area. Calls are now routed out redundant capacity and the issue is resolved. There was no impact to 911 service.