Month: October 1999

Fri Oct 29 12:19:10 PDT 1999 — We’re renumbering the active file on Ultra, aka news.sonic.net, and this means that posts you make right now will be delayed, and will show up in a little while. Additionally, the article counts which you see before you enter the group may be way off from the number of articles that are in the group.

In other news news, our new satellite linked news feed and server are coming together. The satellite dish is being installed on Monday. This new machine has more than 100 gigs of spool space, half a gig of ram, and will be able to do a really good job with all of the Usenet, including binaries. -Dane

Wed Oct 27 12:14:18 PDT 1999 — Focal just turned up more capacity on the ###-9811 lines. This should alleviate busy signals on the 9811 dialup pool. To find a local focal number you can use our pop finder tool at www.sonic.net/cgi-bin/pops.pl -Steve

Mon Oct 25 12:35:56 PDT 1999 — We will be moving our Covad DSL circuit to a new switch tonight. This will happen at around midnight and will cause a short period of downtown for Covad DSL customers. Also at this time we will be adding more memory to bolt.sonic.net so shell access will not be available for 10 to 20 minutes while we do this upgrade. -Steve

Sun Oct 24 15:36:07 PDT 1999 — Slow Usenet and mail server performance. Ultra has been having trouble accepting new posts intermittantly since Thursday. I’ve just restarted innd, and it’s performing well now. We’ve got a new news server on it’s way, so we’ll limp along for another week or few on Ultra as we deploy the new box. The new server, our fourth news server incarnation since Sonic.net’s inception has half a gigabyte of RAM and over 100 gigs of Usenet news spool space. It’s a huge, fire breathing, satellite linked super-fast powerhouse. Notably, it’s also our first non-Linux server, and our first non-INN based Usenet news server. The new system runs the Typhoon Usenet server under Solaris 7. Over 50,000 newsgroups will be available, and binaries will be quite complete. The new satellite downlink allows for two full T1s (3Mbps) worth of inbound bandwidth while freeing space on our terrestrial circuits for customer traffic. We are also expecting an upgrade from 3Mbps to 10Mbps in the next few months to handle expanding Usenet volume.

As a result of Ultra’s trouble receiving new posts, Buzz, our inbound mail server has had a backlog of inbound mail delivery. qmail on Buzz will deliver 10 local messages concurrently, and what was happening was that the news posting backlogs for spam@sonic.net were using up all 10 queue slots for many minutes as a time. We’ve deferred the spam mail to a local mailbox on Buzz for a time, and the queue backlog is clearing. Sorry for the inbound email delay this afternoon! -Dane

Wed Oct 20 23:18:04 PDT 1999 — Tonight a bit past midnight we’ll be replacing the makeshift fuse assembly in the the RedBack DSL router with the new fuse they delivered today. Downtime will be under five minutes. -Dane

Wed Oct 20 02:57:29 PDT 1999 — The trouble with the RedBack DSL router was found to be a blown fuse which serves the packet forwarding engine slot.

The 5 amp fuse had blown, and RedBack support reported that they had found that a 7.5 amp fuse was required. New chassis are shipping with the 7.5 amp fuse, but somehow we didn’t get any info on this, or a new fuse shipped to us.

I built a tiny new fuse assembly by rigging up a 6A configuration using a soldering iron, some wire and two other fuses, and the system is back online. RedBack is rushing us a full set of replacement fuses for this unit. Sorry about the downtime! -Dane

Tue Oct 19 23:37:00 PDT 1999 — The RedBack appears to have a failed packet forwarding engine. The FE is one of the two main CPU cards – the control engine and the forwarding engine work together as the core of the SMS500. I’m awaiting a call back from RedBack’s 24 hour support group for confirmation. If the hardware is toasted, they’ll have to drive a replacement card up here, or I’ll have to go pick it up in the South Bay. Either way, I hope to get back online by 5AM at the latest. Sorry about the downtime! The unit is configured with dual redundant power supplies, and we’ve invested in a 24x7x365x4hr response support contract, so we have done what we can to get very quick response from RedBack in case of this type of failure. I’ll make an update here in the MOTD when I have more information. -Dane

Tue Oct 19 22:56:34 PDT 1999 — The DSL gateway router that serves our PacBell DSL customers has gone off the network. It’s still on, and reachable via serial, but it’s Internetworking is down. I’m on my way to the office to reboot it now. It should be online by 23:10. I’ve got a call in to RedBack to see if they know what may have happened. The OS was upgraded a few days ago, so perhaps the new release has some bug that we haven’t seen yet. -Dane

Mon Oct 18 10:31:53 PDT 1999 — We are upgrading the OS in our RedBack DSL router right now to a new version. PacBell DSL customers may notice a 30 second pause in Internet connectivity while the machine reboots. This new release may address issues where some customers are having connectivity problems which are then resolved by Sonic.net clearing their session in the RedBack. -Dane, Eli, Dave

Mon Oct 18 14:51:36 PDT 1999 — SBC/PacBell has announced a major $6 billion broadband initiative. This will mean much greater availability of ADSL service for customers who are further from central offices and who are in cities which are not currently being served with DSL. We’re very excited about the expansion of availability of ADSL to many more of our Sonic.net customers.

For full info, see SBC/PacBell’s press release dated October 18th at the following URL:

www.sbc.com/News_Center/Home.html