- Moved the Pasadena office web site root directory down one level. This
rearrangement will allow for us to host multiple sites on our server.
- Assisted Stan Silverman with configuring a third Squid server for the
Menlo Park web site.
- Finished the Solaris install on Baldur. Installed Apache, JDK, Sendmail,
and Perl on it. Set up accounts for tnw and isti.
- Set up for accel+proxy on all Squid servers. This is the workaround for
the Internet Explorer persistent connections bug. This allows the server
to correctly serve all requests for our site. It will only return a 403
error if the browser sends a request for a URL on another web site.
- Upgraded Squid to 2.3-stable4 on all Squids.
- Hacked on the EqInTheNews scripts. Moved all but the two CGI scripts out
of the web server space for security reasons. Had to modify all the
scripts to reflect the move.
- Enabled softupdates on Ehzsquidpas.
- Talked to Jill McCarthy about the NEIC web server proposal.
- Talked to Vickie Clark in Golden about web server issues.
- Upgraded Htdig on Agent86 to 3.20b3.
- Fixed Mandy's pager number on the paging web pages.
- Investigated some email problems reported by Sue Perry.
- The Squid server in Reston was logging bad block errors over the weekend.
Built a replacement disk from a recent backup and shipped it to Judy
Konnert. She swapped the disks on Tuesday. This fixed the disk problem,
but the machine reported a memory self-test error on reboot. Did some
testing, and it appears to run all right. She disabled the memory
self-test for now.
- The third Squid server for Menlo Park went live on Tuesday. It is located
in Reston. Its node name is Sqehzeast.
- The bad Seagate 9GB drive from the Spring temporary wavepool turned out
to still be under warranty. Sent it back to Seagate under RMA 5336063.
- Had a SCSI controller that caused system crashes. Sent it back to Adaptec
under RMA RP01029930.
- Set up apache on Baldur.
- Listened in on a conference call about Electronic Timesheets from 10:00
to 11:00 on Tuesday.
- The Reston Squid server was logging '403' errors. This turned out to be
caused by the 'eqwebback' name not being in DNS. It was accidentally
deleted by someone in Menlo Park. Charlene fixed it on Wednesday.
- Started doing garbage collection on the 66 subnet address space.
- M3.5 near Yorba Linda at 04:50 on Friday. Used the eqnews scripts to
create a special page for it.
- The new large-format plotter arrived. Spent most of Friday afternoon
assembling it.
- Fixed the time zone on Sqehzeast. Had to restart Squid and cron so that
everything would work correctly.
- Set up automatic start of the TriNet Watch programs on Baldur.
- Stan Silverman implemented softupdates on the ehzsquidmenlo Squid server
on Monday.
- Sent the old disk from ehzsquideast back to IBM under RMA 60727358.
- Put 512MB of memory and an additional CPU in Iron and Rtdev.
- Set up a virtual server to act as www.relm.org. Sent Charlene the
information to set up DNS for this. Submitted the DNS change to Network
Solutions.
- Began building a new system disk for the usgs-squid hardware upgrade.
- Did an outreach from 11-2 on Thursday at Cal State L.A.
- Got the replacement SCSI controller back from Adaptec on Friday.
- Set up a test CGI script for the duty seismologist to use for adding
special reports to the main Pasadena web page.
- Assembled the hardware for the new upgraded Pasadena Squid server. This
is an Athlon-1.2, which should increase our web service capacity
significantly.
- There were problems with event notification email to eqalarm_local on Jet.
The messages were being delayed, and were arriving 45-50 minutes later than
the mail to the public mailing lists. This turned out to have been caused
by the peculiarities of Sendmail and its interaction with the GPS mail
server. Set up a new mailing list on Eqinfo to replace the eqalarm_local
alias on Jet. This fixed the problem.
- Got back the replacement for the IBM 9GB disk sent back on 4/16.
- Researched DNS and Sendmail issues to come up with a plan for Eqinfo to
resolve to multiple machines for redundancy.
- Installed 'zip' on Willow for Bruce.
- Talked to Jill McCarthy in Reston about possibly signing NEIC up with
Akamai to solve their content delivery problems.
- The new Pasadena Squid server has one problem. The Abit main board does
not have the capability of being set to automatically boot after a power
failure. We are going to return it and replace it with a Soyo K7VTA-Pro,
which does have this capability.
- Ojai crashed with an apparent hardware fault on Monday, Tuesday, and
again on Wednesday. After the Wednesday crash, opened the machine up to
check, and found the CPU heat sink was uncomfortably hot to the touch. The
fan was working, but the air intake on the front of the machine was clogged
with dust. After cleaning, the air flow was greatly improved.
- Attended the CUBE users meeting on Friday morning.
- Cleaned up the old manuals and such that were left over from the shelves
in Phil's old office.
***
- Enabled several performance enhancements on the Earthquake Hazards
Program Squid servers. Also upgraded the USGS Pasadena Squid server
to a faster machine.
- Set up the new large-format plotter.
- Set up DNS, web, and mail services for the RELM project.