- The Caltech electrical shop switched the non-generator circuits in
535 on Monday. They are now fed by the new electrical service panel on the
north side of the house.
- Started building up the new Ultra-60 that will be the hot backup
for Rtdev/Magma. This machine is going to be called Craton, and will
be at 131.215.65.89.
- The Airtouch SNPP server was down from about 09:00 to 10:00 on Monday.
- The Caltech electrical shop switched the generator circuits in 535 to
the new power feed on Tuesday. After we asked nicely, they stayed a
little late to be sure that the generator and transfer switch were
operational on those circuits. Full emergency power functionality was
in place at about 17:15.
- Los Nettos was down from about 06:00 to 10:00 on Tuesday. Campus DNS
was unavailable from about 08:30 to 10:13.
- Located a computer recycler that will take our dead monitors without
requiring us to pay a surcharge. Their URL is www.siliconsalvage.com
- Genie died at 17:41 on Tuesday. It did not reboot. Came in about
21:00 and hooked a terminal up to its console port. The machine was
hung. After a power cycle, it rebooted normally.
- Genied rebooted itself twice over Wednesday night. Examination of the
logs showed it complaining about memory errors in the SIMMs in U0403 and
U0503. Reshuffled the SIMMs and rebooted. Errors persisted. Moved the
SIMMs to a different set of slots. The machine crashed again at 09:11 on
Friday, complaining about memory errors in the new locations. Removed the
bank of SIMMs and rebooted.
- Started cleaning up my office and rounding up dead equipment to take
to the recycler.
- Added a redirect to the Terra10 web page so that all requests will be
redirected to Rift.
- There was a problem with the waveforms on the event review page on Rift
on Tuesday morning. This turned out to be caused by a collision between
Doug's copy routines and the automatic rsync I had set up as an interim
measure. Rsync was deleting files it should not have been deleting.
Since Terra10 is no longer the official internal web server, I turned off
the rsync, and this fixed the problem.
- Decommissioned Terra10.
- Made accounts for small and rtem on Craton.
- Jill McCarthy called with some questions about the information I had
sent to Akamai in their EdgeSuite Discovery Document. She is having to
rewrite her requisition for their service.
- Fixed up the Trinet internal web page on Rift to clean up obsolete
links and missing "/" characters.
- Kimo arranged for a room in the basement of S. Mudd to be used as a
staging area for the stuff we are taking to the recycler. Moved the
dead monitors out of the basement storeroom.
- Attended the Caltech Property Services meeting.
- Added Baldur to Big Brother monitoring.
- Counted up the junk for recycling.
- Joe Monaly from ITS hooked up the network ports in the attic of 535 and
in the new shop behind 535.
- DNS was broken from about 07:00 to 14:00 on Tuesday. The details are at
http://bort.gps.caltech.edu/stan/mail-archive/msg00030.html
- Attended Katrin's going-away luncheon on Wednesday.
- Moved my office to the back office in 535.
- Mail on Bigone was broken on Wednesday. There was a spam that caused
the mail symbiont to crash, and the repeated mail crashes created so
many core dump files in sys$system that the system disk filled up.
- The Code Red Worm came through this week. On the USGS network, it
infected five machines. ITS blocked these five machines at the border
routers.
- The Big Brother HTTP test failed for the Squid servers in Menlo Park
and Reston on Friday afternoon at 15:30. The machines are up and
serving traffic, but Big Brother can't see them.
- Called Bob Logan on Saturday to have the block removed from the
five Windows machines. Bob Dollar fixed them on Friday night.
- The HTTP transport problem was still happening on Monday morning.
Stan Silverman learned that it was caused by DREN blocking port 80
to cut off all web access to the USGS through their network. This
had the effect of cutting off about 2/3 of our normal web server
traffic.
- Jet died at about 20:02 on Monday. It failed to reboot. Came
in about 21:30 to fix it. It was complaining about file system
damage, and also that the 2GB of RAM in Board 0 Bank 0 was bad.
Was able to get it running again, but could not get the memory
to self-test correctly. Left it running like that overnight.
- Did surgery on Jet on Tuesday morning. Reseated the memory on
Board 0 and was able to get it to pass self-test.
- Attended a meeting to discuss disk space for the databases.
- Recycling trip to Silicon Salvage. They pulled a bait-and-switch.
On the phone they had said that they would take monitors, but when
I got there, they didn't want to take them unless I paid them $50.
I paid just to get rid of the load, but I was left with the feeling
that one gets after buying a used car.
- Added Raven and Mark Simons' cell phones to the eqpager mailing list.
- Researched the 3ware RAID cards mentioned in the article about building
a cheap 1TB file server.
- Spring was reporting high iowait levels. The disk at target 3 in the
temporary wavepool was failing. It was not reporting errors, but it was
reporting 80% busy. Replaced it with a spare, and the activity on it
was reported at only 40%. Went to the Seagate web site and got
RMA 5384857 for it.
- Sent in a correction to my account number for travel reimbursement
for the trip to Golden last November.
- Packed the space around Jet in the rack in an attempt to force more
air through the system chassis to improve cooling. This lowered the
reported internal temperatures by about 2-3 degrees.
***
- Set up a new machine as the Trinet Intranet web server. The new
machine is a Sun Ultra-60, which is considerably more powerful than
the old server. It serves earthquake information for internal use,
as well as being the base for the duty seismologist for reviewing
events.
- Collected all of the old computer junk to take to a recycler.