July 25th, 2010
vis.nu Networks was down for an extended period of time this weekend for two reasons:
- The Washington Building, the building which houses vis.nu, had a chunk of its electrical infrastructure replaced during a nine hour downtime this weekend. We were given hours of notice.
- During that time, the drive in our Phoenix server, which still houses the DNS for our systems, failed. Since our usual reporting systems were offline, this wasn’t noticed until a couple of hours ago.
I’ve put a bandaid in place to handle the DNS, but systems that run in Phoenix are offline and will remain offline for some time. No ETR. More as I know.
Posted in Uncategorized | No Comments »
March 30th, 2010
Dear Tacoma Telematics Clients,
Tomorrow, Wednesday 03/31/2010, Tacoma Telematics will experience a
down time of approximately six hours. The power company is updating
components of the electrical system of the historic building where our main
server is located.
In order to insure no loss of data, please complete and save any work prior to
6:00 PM PDST (9:00 PM Eastern). We expect to resume services by 12:00 AM
PDST Thursday, April 1 (3:00 AM Eastern).
We sincerely appreciate your patience and understanding regarding this
upgrade to our electrical system...
It goes on from there, but that’s the important part.
Posted in vis.nu Networks | No Comments »
March 2nd, 2010
As of right now, the old indra is off on an unrelated IP, and a new DNS only server is sitting on the old vis.nu IP. The move is complete. The only thing remaining is to get the integrated Mission Control back up, but unfortunately, that’ll have to wait a little while so I can attend to other pressing matters.
Until then, here are links to management systems:
Posted in Uncategorized | No Comments »
February 23rd, 2010
The main vis.nu systems are moving from Arizona to Washington this week, and there’s a lot going on. If you ‘re looking at this post from the MOTD, I strongly encourage you to to go the website and read further. Aside from the most recent update, I’ll keep everything behind a break for brevity.
Update: 2/25 04:32: On the old system, I’ve turned postfix into a secondary-only backup MX which points at the new server, dovecot and apache have been shut down. Unless you need access to files on the old system, you should be completely pointed at the new system at this time. A placeholder web server which instructs browsers that the system has moved and their DNS is stale is in place.
Read the rest of this entry »
Posted in Washington Move, vis.nu Networks | No Comments »
January 16th, 2010
The database server on vis.nu has been shut down, and all connections are being forwarded to the new database server at the new location
To recap, at this point, mail is being saved on the new server and is active via NFS, and all database connections are being forwarded to the new MySQL server. The only remaining active services are shell and web, and movement of those services are underway.
Posted in Washington Move, vis.nu Networks | No Comments »
January 15th, 2010
As of about 30 minutes ago, vis.nu mail is back up and running.
All active domain maildirs were copied to the new server, and then the new server’s mail storage partition was NFS mounted here via the VPN, so all mail is actually being stored on the new server.
This means mail actions might be a bit slow for the next few days as I continue to move other services to the new server. The plan is to move things a service at a time so the switch over is nearly seamless. Databases are next, followed by hom directories and a standard shell environment.
The web directories (after some significant cleaning out) are last, and once that’s complete, I’ll make the DNS changes to point to the new server, move this machine off this IP, and move the DNS-only VM into place.
If you’ve got a domain that points to vis.nu, you’ll want to change to the new IP, 69.10.218.166. I’ll be contacting people individually about that later on.
Posted in Washington Move, vis.nu Networks | No Comments »
January 14th, 2010
Mail service will be shut down from 7 PM to 9 PM as I take a snapshot for migration to the new server in Washington. This includes all the mailing lists. I would have picked a better time but I’m starting to be under the gun.
Update 9:37 PM: I lied– I sometimes forget how long tarring up 10G of maildirs takes. I’m almost through it, but the largest accounts are still to compresss.
Update 1:08 AM: The compression portion is done, and the transfer is about 75% complete. After the untar, the system will be back up and running, but at somewhat reduced bandwidth until the transfer is complete. More on that later.
Posted in Washington Move, vis.nu Networks | No Comments »
November 4th, 2009
Hell, this is what? Move #70 it seems like, not to mention that this is state #3.
vis.nu is over a decade old, and while it hasn’t seen an upgrade in a while, I haven’t ignored it. I’ve just been very busy, as some of you know. The new server (brahman) is a machine given over to Xen entirely for vis.nu. This means that major functions have have their own virtual machines. The shell server, the one most of you will see, will remain indra, while we’ll have a return of vac and kali and the introduction of a few others.
The first thing to move, mail services, will essentially be transparent as it’s completely separated from the shell services anyway. There may be the usual DNS lag, as well as some short downtime while I move the actual stored mail from Arizona to Washington.
After mail, we’ll be moving web, after web, we’ll be moving shell services (which I think only four of us use anyway).
If you have any questions, as usual, email me.
Posted in Washington Move, vis.nu Networks | No Comments »
September 14th, 2009
We were down for a few hours, still don’t know why, but we’re back up now and the move continues apace.
Posted in Uncategorized | No Comments »
May 22nd, 2009
The RPB in the AZ colo failed, and the RTC on indra failed to come back up properly. Good thing we’re moving soon, the machine we’re on isn’t going to be good as a server for much longer.
Everything appears to be running normally. Let me know if you run into any further issues.
It should be noted that Thursday was the one year uptime mark for this incarnation of the server. This is second place after the 700Mhz Athlon Slackware vac.vis.nu which had remained up for 17 months.
Posted in Uncategorized | No Comments »