Month: June 2000

We tracked the slowdowns for web and mail…

Sun Jun 11 11:34:49 PDT 2000 — We tracked the slowdowns for web and mail back to the file server, ‘freezer’. It turns out that the Ethernet connection had misnegotiated: ‘freezer’ was running 100 megabit full-duplex, but the switch was only running 100 megabit half-duplex. Hard-setting full-duplex on the switch solved the problem. -Scott and Dane

Network Appliance migration complete.

Sun Jun 11 10:21:10 PDT 2000 — Network Appliance migration complete. We’ve completed our migration of data from our old NetApp — ‘fridge’ — to our new NetApp, ‘freezer’. Some folks have reported intermittent slowness with some web pages, which we are now profiling and trying to pin down. -Scott and Kelsey

More on NetApp upgrade.

Sat Jun 10 22:42:19 PDT 2000 — More on NetApp upgrade. We have seriously underestimated the amount of time to migrate web and mail space to the new NetApp. New estimates based on timing of full mirrorings (including deleting files that no longer exist on fridge): /nfs/www/realaudio 2 minutes /nfs/www/WWW-data: 5 minutes /nfs/www/httpd: 15 minutes /nfs/mail: 2 hours /nfs/www/WWW_pages: 4 hours These estimates are based on timing our continuous updates since Wednesday. Times are not cumulative, as we will be running these updates in parallel. Based on that, here is the upgrade schedule: From 1am-3am: /nfs/mail will be read-only. This means you will be able to read and send email, but will not be able to delete messages from your mailbox. Incoming mail will be queued on our local servers and delivered at 3am. From 1am-5am: /nfs/www will be read-only. Web pages can be accessed, but you won’t be able to make any changes. Further, cgi-bins that write data to files under /nfs/www won’t be able to do so. We apologize for misestimating the amount of time to complete this upgrade. These are also based on ‘worse-case’ scenarios, so downtime will probably be less than we estimate. -Dane,Scott,Kelsey

Dane Jasper, founder and CEO of was

Fri Jun 9 18:07:46 PDT 2000 — Dane Jasper, founder and CEO of was recently interviewed by The Business Journal, and we’ve made a copy available on our home page. Dane talks about our six years as the leading local Internet provider, plus’s growth and future. See the ‘What’s New’ section of our home page to read the interview.

This Sunday morning after 12:30AM, we will be

Fri Jun 9 18:03:48 PDT 2000 — This Sunday morning after 12:30AM, we will be upgrading and rebooting our core DSL router. This will result in three to five minutes of downtime for PacBell DSL and Broadlink DSL customers, as well as T1 customers connected via FR/ATM.

Next, we’ll be shutting down our new NetApp filer, to move our 1.1 Terabyte tape changer to it’s SCSI bus for backups. This will pause FTP availability for two to three minutes.

Then, we’ll be migrating user email and web data from the current NetApp filer to the new filer. This will result in read-only availability of web and mail space for 5-10 minutes each, then downtime for one to two minutes while the NFS mount points are moved.

Scott and Kelsey will be doing the work on Sunday morning. -Dane

We’ve completed the migration to piggy; all…

Thu Jun 8 18:04:37 PDT 2000 — We’ve completed the migration to piggy; all multihomed websites are now load balanced and enjoy the reliability and performance of redundant web servers. If you have any questions about the new web server configuration and how this may affect your website please direct them to the newsgroup where we will be happy to answer your questions. -Kelsey and Scott

We’ve completed the migration of all SMTP,…

Tue Jun 6 18:09:53 PDT 2000 — We’ve completed the migration of all SMTP, POP3 and HTTP for to the new Alteon AD3, ‘piggy’. Piggy’s configuration enables us to load balance all multihome domains. We’ve migrated one class-c ( of multihomes on piggy to test the new load balanced configuration. If you’re domain is hosted in the class-c and you run into problems with it please let us know as soon as possible. We expect that testing will be finished soon and that all multihomes will be load balanced by Friday.

Once the configuration changes have converged and we have migrated all of the multihomes to piggy we will setup piggy and kermit (the other AD3) in an active-active redundant configuration, thus insuring no single point of failure within our server LAN. -Kelsey

We’ve completed the migration of FTP data to…

Mon Jun 5 21:58:41 PDT 2000 — We’ve completed the migration of FTP data to the new Network Appliance network filer. The new NetApp F740, ‘’ is to replace ‘’ with 6x the speed, twice as much RAM and much more space. It’s scalable to 900 gigs of fully redundant RAID storage, so we’ve got a lot of room to grow. The fast new fiber channel arbitrated loop drives support multiple filer clustering for high availability.

We’ll be migrating mail, web and shell home directories in the next couple weeks to complete this project. The current schedule is to move mail and web data this coming Sunday morning, June 11th after around 1am.

‘fridge’ has served us very well, with amazing reliability and zero-downtime file-system expansion which has been a unique and wonderful solution for our growing storage needs. -Scott, Dane and Kelsey

New Network Appliance migration.

Sat Jun 3 19:49:04 PDT 2000 — New Network Appliance migration. Tomorrow morning (June 4th) at 1am, the ftp volume will be read-only for approximately 10 minutes while we transfer the last of the data to the new NetApp. After that, we will be changing all ftp mount points over to the new NetApp. The migration should be finished by 1:30am. -Scott and Dane

The Covad DSL backhaul issue was resolved…

Fri Jun 2 21:41:45 PDT 2000 — The Covad DSL backhaul issue was resolved from a performance standpoint at ~17:00hrs, and Covad’s engineers are still trying to determine what caused the problem. I waited to update the MOTD for that resolution, and at this point I will direct further updates to the sonic.dsl Newsgroup. — Eli