- Repartitioned the system disks on Jet and Spring. This was to divide the
swap partition from a single 4GB slice to two 2GB slices. This was because
research revealed that Solaris is not capable of dealing with any one swap
space larger than 2GB. This reveals the 32-bit heritage of Solaris.
- Set up an automatic archiving of the Daily Digital Station Status
messages that the duty operators generate. These are archived on at
http://terra10.gps.caltech.edu/station_dailies.
- Added Jocelyn and Vince to paging.
- Moved the /opt filesystem on Iron from a 9GB disk to one of the new 18GB
disks.
- Fixed the cgi-bin directory permissions on terra10 to fix a problem with
the software defect tracking programs.
- The system log file on Ehzsouth showed some odd messages indicating that
someone was performing security probes on us. Research revealed that the
probes were coming from ORBS, which is an organization which searches for
open email relays in order to stop spam. Someone had apparently nominated
our system for testing as a possible open relay. Checking with the ORBS
database showed that we passed the tests.
- Got a quote from Kingston for 2GB memory for the new Enterprise 450
system.
- ITS upgraded the firmware on the CITnet building switches in the two USGS
houses.
- Set up directories for isti to use for program development on Iron.
- Hector Mine event on Saturday morning.
- Web server traffic after this event was so high that the log files had to
rotated and cleaned out twice a day for several days in order to prevent
the disks from filling up.
- Traffic peaked on both Ehzsouth and trinet at about 7:00 AM on Saturday.
At this time, both servers were getting about 35-40,000 hits an hour.
Ehzsouth was serving about 2GB an hour, and trinet was serving about 120MB
per hour. The big hit was the Commmunity Internet Intensity Map, which was
downloaded for a total of 14GB from Ehzsouth. This was a large fraction of
the total traffic of 17GB for the day.
- Assisted Bob with installing two new disks on Bbear to hold the data
coming in from the Hector Mine sequence.
- Added some new people associated with the Marine Corps to rpage so that
they could get CUBE pages.
- Fixed DECnet proxies on Bbear for cull. This was necessary so that cull
could run on Bbear, since Bigone was bogged down with other processing.
The fix was to copy the NETPROXY.DAT file from Bigone to Bbear.
- Investigated a reported problem with Contel paging to Caltrans District 9
in Bishop. This turned out to be a problem at the GTE end which was
causing the pages to be rejected.
- Fixed the routing tables on the new swordfish router for the microwave
connection to DWP.
- Did some maintenance on the Lexmark printer in the tan house. Installed
a new photoconductor cartridge and a new black toner.
- Installed some patches on Ehzsouth to help with a problem where the web
server was reporting 'too many open files' and was refusing to serve pages.
- Talked to Pat Murphy about address translation issues for our new DOInet
connection.
- Researched options for increasing our web server capacity.
- There was a problem with paging on Bigone. The pager daemon process had
died silently during the night, and pages were backed up in the queue.
- Fixed rpage on Wwolf. It wasn't working, due to the PAGE$TABLE logical
not being set properly.
***
- The USGS and Trinet web servers experienced their highest traffic levels
yet as a result of the Hector Mine earthquake. The USGS server serviced
350,000 hits, totaling over 17GB of data during the first 24 hours . The
Trinet server served about 2GB during the first 24 hours. The majority of
the web server requests were for shake maps and for the Community Internet
Intensity Map.
- Performed several reconfigurations of the VMS system to deal with the
increased data load from the Hector Mine sequence. Added four 9GB disks to
serve as a buffer to hold the incoming data.