- Participated in a conference call to discuss the implementation of the
Recent Earthquakes/US web pages. Also discussed using QDDS to drive
the Earthquake in the News function.
- Recabled the FDDI to go through the Catalyst 5000 in the basement of
525.
- Put the new RecenteqsUS maps on Horst for testing.
- Moved the scweb-north IP address to Horst. This is a full backup copy
of the Pasadena office web pages, housed on a server in Menlo Park.
- The PC guys in Menlo Park resurrected eqinfo2.wr.usgs.gov on Thursday.
Got it set up and configured.
- Transplanted memory from Fang to Agent86. Both servers now have 512MB.
- The group paging function on the Send-A-Page form was broken. Turned
out to have been caused by modifications I made so that the script would
run under 'use strict' in Perl. One misplaced 'my' and it was broken.
- There was a bogus event on the Recent Earthquakes maps on Friday. Added
a hack to req_prescreen.pl to have it discard any events with origin times
in the future. Submitted this patch to Bob Simpson.
- Set up synchronization of the mailing lists between eqinfo0, eqinfo1,
and eqinfo2. The model is that eqinfo1 and eqinfo2 are to be eqivalent
machines, and eqinfo0 is the backup. Subscriptions are synchronized
between all three machines, but manual changes are only propagated from
eqinfo1 to eqinfo0 and eqinfo2. The eqinfo0 and eqinfo1 servers are
here in Pasadena, and eqinfo2 is in Menlo Park. System status is
displayed at http://bort.gps.caltech.edu/mrtg/eqinfo.html
- Put the new memory kit in Hotspot. Turned out that they mistakenly
sent us the wrong kit. Rosemary will call them and arrange to get the
correct one.
- Installed the new DLT tape drive on Iron on Monday.
- Pluton had a disk failure on drive7 card0. Power cycle brought it back.
Talked to Egill about ordering a spare drive for this array.
- Put the Recent Earthquakes/US pages on Graben.
- Added a password for the www.scign.org/minutes directory.
- One of the building switches in 525 failed on Tuesday. Called ITS and
Joe came by and replaced it.
- Had eqinfo2 added into the MX records for eqinfo.gps.caltech.edu.
- Set up MRTG/Big Brother monitoring on Horst and Graben. The status of these
machines can be seen at http://bort.gps.caltech.edu/mrtg-ehp/
- The disk on Pluton failed for good on Wednesday. Got an RMA from Maxtor
and sent it back for warranty replacement. The RMA number is 0200777128.
- Hotspot died at 19:30 on Wednesday. Had to manually fsck the disks to get it
to boot.
- Attended the HTSI Town Hall company meeting on Thursday morning.
- Changed the CIIM on Agent86 to use ssh for transport. This will eliminate the
need for the web server to export filesystems over NFS.
- Sent the carcass of Foreshock to Menlo Park so that Stan Silverman can swap
it out with Horst. Horst has been unstable, crashing whenever it is under heavy
load. Also, processes have been dying with signal 11 errors, which typically
indicates a hardware problem.
- Looked on Freshmeat for Help Desk applications. Found several that were promising.
Set up a test of PHPHelpDesk at http://bort.gps.caltech.edu/phphelpdesk/
- Installed 4GB in Spring, 2GB in Hotspot, and 4GB in Jet. Jet and Spring now
have 8GB each, and Hotspot is at its maximum of 4GB.
- Got jdk1.2b10 built and installed from the ports collection on Bort, Horst, and
Graben. This allows these machines to run the new Java implementation of CNSSM.
- The warranty replacement disk for Pluton arrived on Wednesday. Installed it and
rebuilt the array.
- Installed patch 105669-11 on all the Solaris 2.6 systems. This is a security
patch for CDE libDtSvc.
- Set up recenteqsUS on Horst and Graben.
- Got the carcass of the former Horst back on Thursday. Rebuilt the machine with a
new motherboard to fix the crashing. Resurrected it as Foreshock on Friday.
- There were some problems on Friday night with QDDS and CNSSM dying on Horst
and Graben.
- Martin Luthor King holiday on Monday.
- Jet crashed at about 01:47 on Tuesday. The error had scrolled off the console,
but the machine was complaining about memory errors on Board 0/J3700. Went in and
swapped and reseated all the SIMMs on that board.
- The spare disk for Pluton arrived from buy.com on Tuesday.
- Jet crashed again on Wednesday at 16:05 and 18:20. Hooked it up to
the serial port on Flint on Thursday morning to log the console output.
This was done with Kermit on the serial port. Jet crashed again that
morning, and the error was complaining about memory errors on J3601 on
Board 0. This is one of the new 2GB memory kits. Removed this kit and
rebooted.
- Moved my office to the new location in the yellow house.
***
- Got the mailing list server running on a strategic triad of machines.
Eqinfo1 and Eqinfo2 are the main servers and will load-share. Eqinfo0
is a backup for both of them. Two machines are in Pasadena, and the
third is in Menlo Park.
- Added memory to Jet, Spring and Hotspot. One of the memory kits on
Jet was experiencing errors and causing the system to crash. Contacted
the vendor and received a replacement kit.