- Set up mailing lists for Ned and RELM. The mailing lists will live
in a virtual domain on the Eqinfo mailing list server, and will all
be in the form 'listname@relm.org'.
- Got a copy of Akamai's quote for providing the EdgeSuite service for
NEIC. Talked to Chris Carston at Akamai and asked some questions about
technical details of how it works.
- Unpacked the new [refurbished] Ultra-60 that will be a hot standby for
Magma, Granite, and Rtdev.
- Called ITS to arrange for them to fix the wiring in the attic of 535
so that the refinishing can begin.
- Moved Avalon into the Timers' room and put the CEDAR disks on it.
- Upgraded the C and C++ compilers on Hotspot. This is for a test that
Lupei is doing. Had to get an updated license file before they would
run properly.
- Got back a replacement disk from Seagate to replace the one that was
sent back on April 10th.
- Fixed the NSI auto-startup on Baldur. This was failing because of
some 'stty' commands in the .login file for the tnw account. Removing
this file fixed the problem.
- Found part numbers and ordering information for media for the new
large-format plotter and Linda ordered several rolls.
- Alan Yong had a problem trying to access ftp on Agent86. This turned
out to be caused by the fact that the first 16 addresses in the DHCP
pool had reverse lookups of 'DHCP-66-xx.caltech.edu', which caused
problems for services that are restricted to the gps subdomain. Requested
ITS to fix this in DNS.
- Built up two new system disks for the new Trinet web server and for a
new mailing list server.
- Set up a rule to archive mail to operators with 'Xsomething' in the
subject line. This is for archiving information about Earthworm hubs.
- Set up the mailing lists for relm.org.
- Talked to Chris Carlston at Akamai about the EdgeSuite service. It
appears that this service would work well for us if we can get it for
a palatable price.
- Got back the bad DIMM from Ehzsquideast. The company web page at
http://www.pdpsys.com claims a lifetime warranty on memory, but they
say we have to take it back to Fry's.
- Helped Mike Watkins with a terminal server problem.
- Added Sue Perry to the pager lists.
- Talked to Jill McCarthy about Akamai and web server issues.
- Got the CGI interface for EqInTheNews working. Doug put a button on
the event review page to activate this.
- Configured the new Trinet web server.
- The system clock on Hotspot was off. This was caused by xntp being
hung in startup. Disabled startup and added a crontab entry to just
run ntpdate periodically.
- Boron's clock was off. The root crontab was gone, probably from the
last time it was hacked. Added a crontab entry to run ntpdate.
- Installed the java plugin on Baldur. It almost works, but it hangs
Netscape.
- Alan Yong had more problems with DHCP and DNS. The nameserver files
had not updated properly from the last time they were updated.
- Registered 'shakemap.org' for Dave.
- Replaced a failing cooling fan in one of the disk enclosures on Genie.
- Talked to Jill McCarthy about her proposal for signing the earthquake
program web sites up with Akamai EdgeSuite.
- Installed the DBI and DBD Perl modules on Baldur for use by Trinet Watch.
- Ojai crashed at 01:40 on Sunday morning, and again at 10:15 and 16:07
on Monday. In all cases, the crashes occurred when it was running REFORM.
This points to a probably hardware problem. Since the VMS machines are
slated for decommissioning this year, we decided to work around this problem
by moving the REFORM jobs to run on Carizo. This will reduce the stress on
the Ojai hardware.
- M3.9 event near Devore at 10:13 on Monday. Web traffic peaked at 58/sec
in Pasadena, and 140/sec in Menlo Park.
- Moved Kate's computers into her newly remodeled office.
- There were some problems with the FDDI ring on Tuesday during routine
reboots of Jet and Spring. This turned out to be caused by the "B" port
connection on Ojai being improperly seated. The ring was running in a
wrapped state, so when another machine rebooted, the ring was segmented
into two pieces, causing lost connectivity. Reseated the connection.
- Researched Network Address Translation for allowing access to the
new earthworm system FortScottyHub.
- Got a quote for an Ultra-10 memory upgrade kit and a four-disk array
for Hotspot. Rosemary ordered these on Wednesday, and they arrived on
Friday.
- Added Jamie to the Operators mailing list on Iron.
- Set up Nickel as a desktop machine for Ellen Yu.
- Big Brother reported a swarm of problems late Wednesday evening. They
turned out to be false alarms that were caused by the 65 net interface
on Terra10 going senile and dropping packets. A reboot of the machine
at 01:30 Thursday cleared the problem. Also, moved the primary BBNET
function to Ehzsouth, and made Terra10 the failover machine.
- Walker Pass earthquake swarm on Thursday afternoon.
- Fixed the mail statistics-gathering script on eqinfo on Friday morning.
It should now report accurate message delivery counts.
- Continued configuring the new Trinet web server.
- Installed the new Trinet web server at 131.215.65.167 for testing.
- Put memory in Wacke, and the new disk array on Hotspot. Created
the temporary wavepool area.
- Helped Mandy get NAT working on Lamprey for FortScotty.
- Added a paging function to the new internal web page on 'insider'.
- Set up a link for the HP DesignJet printer status page on the internal
web page.
- One of the power supplies on the Jet AA RAID failed. Called nStor and
got RMA number 26162 for it. Paul sent it back to nStor on Thursday.
- Found a problem with the access restrictions on the Western Earthquake
Hazards Internal Web Page in Menlo Park. The page for sending messages
to belt pagers was open to the world. Notified Stan Silverman and sent
him a fix for the problem.
- The Lexmark color printer in the 535 is complaining with a
'Service Engine Board' error. Talked to Linda, and she will call
service for it. Also, the Lexmark in 525 was low on all supplies.
Replaced all the color toner cartridges and ordered replacements.
- Bruce set up a procedure to copy Shake Map files to the new Trinet web
server. When it ran, it overflowed the maximum allowable connections on
the SSH daemon. Had to set the allowed connections up higher so that it
could run.
- Talked to Ken Ou about ordering a KVM switch for the telemetry room
for the Trinet web server and Squid server.
- Kate had a problem compiling on VMS which turned out to be caused by
a network problem on the Tejon hub. Ken Hudnut's two machines were
plugged into this hub, and they were filling up the network bandwidth.
This was causing Kate's DECnet connect to time out. Moved them to another
hub on the DECswitch.
- Modified the pager.pl CGI script to allow for pager groups defined on
our system. The new script uses an additional file called 'page-groups.list'
which defines the groups of pagers.
- Patrick reported that Magma was seeing dropped packets in the data stream
from Carizo. This turned out to be caused by the fact that it was on the
same hub as Ken Hudnut's two machines. Ken's machines were eating the
network bandwidth. Moved them to yet another hub.
- Replaced all three color toner cartridges on the Optra SC in 525 on
Tuesday.
- Jet reported a CPU usage problem early Wednesday morning. At the same
time, it started reporting excessive CRC memory errors, all on a one SIMM
on CPU board 4. The SIMM repeatedly failed self-test, and when it was
moved to another slot, the problem followed it. Finally, we had to
remove the entire bank of 8 32MB SIMMs to fix the problem. This left
Jet running with only 3840MB of physical RAM while we look into replacing
the failed memory.
- Worked with Mandy to finish up the NAT tests and set it up for the
Fort Scotty Hub deployment.
- Helped with the Fort Scotty Hub deployment on Thursday and Friday.
- Moved the Lamprey net connection from the 65 net to 66 on Thursday.
- Got a four-port KVM switch from Ken Ou and put it on the PC in the
telemetry room.
- Called Radiant Resources for quotes on a replacement memory kit for
Jet and for a NIC for the new Ultra 60.
***
- Set up a new internal web page for the USGS Pasadena Office. Also set
up the EqInTheNews scripts to allow the duty seismologist to automatically
create special web pages for significant earthquakes.
- Worked with Mandy Johnson to get Network Address Translation working on
the Lamprey router so that the planned Fort Scotty earthworm hub will be
accessible for remote management.
- Installed a four-disk array on Hotspot as a temporary wavepool area.
This makes the Hotspot configuration equivalent to Jet and Spring, and
has fixed the system problems that were caused by disk I/O overload.