- Finished Jury Duty this week.
- Got the Fortress UPS attached to the USGS web server to talk to the UPS
monitoring software. It turned out that the data port on the UPS has a
non-standard pinout, so it was necessary to make a custom cable to make it
work. The cable is DB-9 male to DB-25 male. The DB-9 end is on the UPS,
and the DB-25 end goes on the serial port on a Sun box. The cable has only
three pins.
The mapping is:
DB-25 DB-9
pin 2 pin 1
pin 3 pin 2
pin 7 pin 4
The connectors are like:
------------------
\x /
\______________/
x is pin 1.
- Set up automatic UPS status checking on the web server. This will send a
page if the UPS detects a power failure.
- Assisted Bob Dollar with finding a way to recover from a lost
administrator's password on a Windows NT system.
- The tape drive on Iron has been flaky lately, so Phil suggested replacing
it. Got quotes on some possible replacements.
- Assisted Scott Lydeen with some email problems.
- Packed up the old Aladdin to send back to Sun. This machine was replaced
by a new Ultra 5 on a trade-in program.
- Got a quote for a DLT drive for Iron.
- There was a power glitch at about 08:00 on Wednesday. This caused some
machines in the two houses to crash.
- Ojai was one of the machines that crashed. The machine was placed on a
small UPS before rebooting, so it should be protected from minor power
problems in the future.
- Set up qpage to work in Iron and the other Trinet development machines.
- Got some information about the Sun Ultra-60 as a possible hardware
upgrade for Iron.
- Assisted Bob Dollar with debugging the ppp connection for the new Duty
laptop.
- Changed pointers for the printers in 250 South Mudd, as Kimo changed
these to spool off of Earth, rather than Kipuka.
- Did some research on cables for the console ports for the headless
workstations in the Jet rack. Made a cable to hook up a serial console for
these machines for use in case of system problems.
- Set up a test email gateway for doing paging through email on Ehzsouth.
- Put these status reports on a web page.
- Fixed the colors on the Lexmark printer in the tan house. Somehow, the
color menu settings on the printer had been changed, which made the overall
color balance different from the other Lexmark printers.
- Assisted Doug with investigating a problem with ALARM dying on Carizo.
This happened with an `insufficient dynamic memory' error. Research
indicated that this was most likely caused by an too-low setting of the
system parameter CTLPAGES. We increased this value and rebooted all
real-time systems.
- Moved node Water back into the Timers' Room.
- There were some mail problems on Bigone after some events over the
weekend. The mail queues were backed up to the point that mail was taking
over an hour to be sent. This ultimately turned out to have been caused by
mail addressed to a machine at unr.edu. This machine had gone down, so all
mail sent to it was backed up in the queue, and this clogged the mail pipe
for all smtp mail. The fix was to clean out the queue and remove the unr
address from the distribution lists. The administrators at unr are not
sure what happened to the machine, so we may resume sending mail there in
the future.
- Installed a new disk for Sue Hough.
- Got mpack/munpack for Solaris to unpack Mime-encoded mail with munged
headers. Normally, metamail or Netscape mail can decode these messages if
the headers are intact, but some mailers disturb the headers so that these
programs can't deal with the attachments. These new utilities enable us to
decode these messages.
- Assisted Joe Franck with debugging a cron procedure for updating the
station status web page.
- Installed the new DLT tape drive on Iron.
- Set up the new Ultra-60 to replace Iron. Did Solaris installation and
configuration on the new machine.
- Took the old Iron and gave it to Bruce Worden as node name Garnet.
***
- Got the UPS attached to the USGS web server to talk to the system so that
the system can page when a power failure is detected.
- Investigated delays with the automatically-generated emails from the
online system. This turned out to have been caused by a machine at UNR that
was down.