OASIS98: BASE Messages: 18 Entries..

Restart of nc_server / adams  5 Aug, 98 ~15:30

The ingest processes locked up at about 15:06.
Reboots attempted including power cycle of fiber-optic xcvr in base.
Restarted nc_server and ingest.  Rebooted Adams.  Everything back
online except...
Cause:  The UW Sonic #2.  (See associated log message)
	It appears that a spurious control character from the sonic
	may have caused a odd sort of hang on the adam net.

Aug 4, 15:00

Closed nc_server covar file via:  nc_close aster

Removed covar files:	$ASTER/projects/OASIS98/results/covar/...
	oasis.980801.nc
	oasis.980802.nc
	oasis.980803.nc
	oasis.980804.nc

Rerunning covars:  	$ASTER/projects/OASIS98/scripts/covar_redo
	for these days and the OKMN_process as well.

EDITED: okmn.q	to add netcdf fields from OU data:
		RH.1.5m.2	HMP35 in Beehive
		 T.1.5m.2
		RH.1.5m.3	Rotronics in Beehive
		 T.1.5m.3

Purpose: cross check of the bowen ration 'sniffer' system by
	looking at the other two t/rh readings being monitored
	by the OU system.

I just changed the cal_file for nuw2 to reflect flipping upside down.
To use this updated file, I've restarted covar on ragwort.
Hopefully, the wind directions ("dir") should look better now.

We forgot to mention it earlier, but the OG&E man pulled our electric meter
out today to check if we had tampered with it.  (The security tab wasn't
on the meter for some reason.)  I think he was satisfied.

Anyway, we decided to let the entire system take the hit as a test and 
took no special action.  Of course, the base computers stayed up on the
UPS (which showed running on battery and later charging the battery.)
The adams all went down and then brought themselves back up with no
intervention by us (including ragwort). 

The power was only off for about 5 seconds at 10:05.

I've just done a system backup to tape#6 by:

cd /usr/local/adm/backup
./backup 6

The next backup should be to tape #1.

Xsun locked up, so I didn't get a chance to review today's cockpits.
I assume all was normal by the way they've just come back up.

I've commented out the psi-band stuff, since I don't expect to install it
here.  Thus cockpit shouldn't have any "R.I.P"'s anymore.

I also have restarted covar on ragwort (data) to despike the UW data.

I've just backed up data from 7/10-7/18.  Disk usage has changed from
80% to 9%.  This was a bit early, but the disk has been filling faster
now that we are running 5 sonics, and I saw no reason to push into the
90% range.  Besides, I plan to do another backup when John and I overlap
(probably just before I leave here) to show him how its done.

I just downloaded the following from the UPS's serial port 
(settings: 1200 baud, 8bits, no parity) at about 11:20 CDT Jul 16.

It says that there were several instances of alarms, but that last
night's event was:

18:11 Inverter Brown Out (code B)
19:00 Near low battery; low run time left
19:06 Low battery
19:25 Inverter Line Loss (code L)
19:29 Near low battery; low run time left
19:33 Low battery

The times don't agree exactly, but seem to concur with our idea that
there were two power outages last night.

------------------------------------------------------------------------

=>date
Apr 16, 21:10:22

=>log
Inverter log - Apr 16, 21:10:24

Date  Start HH:MM:SS Typ Active*

04/16 05:15 00:07:49  L
04/16 04:01 00:55:03  B
04/08 13:33 00:00:21  L
03/21 03:26 00:03:54  L
03/19 20:55 00:02:02  C
03/18 20:51 00:00:01  R
03/18 20:23 00:28:05  L
03/16 23:33 00:02:02  C
03/15 23:27 00:00:01  R
03/15 23:27 00:00:22  L
03/15 23:27 00:00:01  R
03/11 23:06 00:00:32  L
03/11 16:46 00:00:21  L
03/04 17:37 00:00:21  F
03/02 00:17 00:00:23  L
03/01 21:41 00:02:02  C

Alarm Log - Apr 16, 21:10:28

Date  Start HH:MM:SS Typ Active*

04/16 05:23 00:30:07  A
04/16 05:19 00:04:26  D
04/16 05:19 00:04:26  B
04/16 04:56 00:02:36  A
04/16 04:50 00:06:11  D
04/16 04:50 00:06:11  B
02/02 12:40 01:03:03  A
02/02 12:34 00:06:05  D
02/02 12:32 00:07:50  B
01/27 19:26 00:00:01  B
01/27 19:26 00:17:13  M
01/26 19:16 00:01:01  A
01/26 19:16 00:00:14  D
01/26 19:16 00:00:16  B
01/26 18:29 00:01:01  A
01/26 18:29 00:00:15  D

=>ahelp
Alarm condition quick-reference list
(* Indicates active alarm)

* Alarm condition  Ltr Bp Code
- ---------------- --- -------
  Low Battery       A  ._
  Near Low Battery  B  _...
  High Battery      C  _._.
  Lo Run Time Left  D  _..
  Low AC Output     E  .
  High AC Output    F  .._.
  Output Overload   G  __.
  Hi Ambient Temp   H  ....
  Sink Overtemp     I  ..
  User Test Alarm   J  .___
  Reserved Alarm    K  _._
  Check Cooling     L  ._..
  Check Battery     M  __
  Check Inverter    N  _.
  Memory Check      O  ___
  Shutdn Activated  P  .__.

=>

When aster's display is locked up, the following comands from Gordon will
revive it, without having to reboot:

(on an aster window on cocklebur)
ps -ef | grep Xsun
[get pid]
su
[enter password]
kill -9 [pid]

(must log in again to aster)

  Raw data files, up to 7/3/98, have been ftped to
stout.

If you make changes to okmn.q and want to regenerate the data, do:

	cd results/OKMN
	okmn_process 980624.ast

  Data from 6/17 to 6/29 has been backed up on
tape. These files have also been transferred to
stout, directory "/net/aster/projects/OASIS98/raw_data/all",
via ftp.

Because I added Tsoil.(1,2) and Gsoil.(1,2) to okmn.q yesterday, 
I changed covar.config to include these as common variables between
NCAR and OKMN.  

Restarted covars.

  While backing up files to tape, we ran out of
disk space! The archive tasks shutdown at 14:44 (local).
The tape backup shows only one large file written
to tape. To play it safe, we decided to ftp the files
to stout at directory /net/aster/projects/OASIS98/raw_data/all.
The files were transferred as binary.

  Software changes have been made to sonic_atik.c and
the "did" of the level sensors.
  "sonic_atik.c" was changed to handle the new level
sensors on the ATIs. According to ATI, the transfer
function is:	x degrees = counts * 1.5 / 1024
"sonic_atik.c" was changed to use this as the default
gain setting. If you need to change this to a different
value it can be done in "prep.config". The function
"sonic_atik.c" can accept 5 input parameters. Parameter
5 is used for this gain term.
NOTE: you must provide a value for parameter 4.
example:  c=sonic_atik("ati.9m.OASIS98,1e-5,2.5,"flag",0,.1)
	 where .1 is the gain value

  The "DID" names for the level sensors were changed from
"alev" and "blev" to "lev.u" and "lev.v", respectfully.
This change was made in prep.config and covar.config.

  Reboot ASTER due to lockup on network link.

Here are some issues with respect to the network link. The
link was setup using the AnyTwist box (refer to drawing on
bulletin board.

(1) The Transceiver has 4 lights to look at. On top of the
unit are 3 lights: green (our network), amber (transceiver link),
and green (power). There are symbols above the lights which give
you a clue as to what they represent. The network and transceiver lights
should be blinking. The power light should be on. There is a 4th light in the
back of the unit which is amber and represents the antenna connection.

(2) If the network link goes down you should check these lights.
If our network light is out then it is our problem. If any of the other lights
go bad then call Harolda at 325-1555 and provide her with any information; ie,
light status, etc. She will pass on the informaion to the necessary people
to get things going again.

(3) IP addresses:

	aster         - 129.15.88.230
	cocklebur     - 129.15.88.232
	defaultrouter - 129.15.88.1

Several reboots of aster and cocklebur this morning associated with
attempts to configure the wireless network connection to OU.