CASES97: Site base Messages, 15 Entries..

Time of the Celiometer PC is off.  It is
1-minute slow.

Restarted ASTER data archival at 1406 GMT.  Archiving quit around 1000 GMT
because the disk was full.

ASTER data ingest died 97/05/08 at ~7:30Z
Caused by the /data disk filling to 100%.
This locked up everything which required a reboot of
ASTER and Cocklebur.

Restarted ADAMS / Archiving at about 14:30Z
after purging some old files.

Matt staked the base trailer down this morning.

I just cleaned the mirror of the all-sky camera today, after being
prompted by a call from Chuck Long.  (I was going to do it today anyway!)

I just did a backup, requesting 4/25-4/28, but it did 4/22 - 4/24 also,
since I never did a bkfiles_kill from the last time.  I just did this
and reduced the disk space usage from 62% to 9%.

This done in an attempt to get platypus treating other hosts properly - didn't
work, so it is still a bit sick (but functional).

Of course, we also had to reboot marigold and cosmos.  (Marigold had died
last night at 1830 anyway).

Here's how to use the new surf routine in S+

	
From the shell, do:

	set_project
		the new version of set_project sets two new environ 
		variables: 
			SURF_NETCDF_DIR
			SURF_NETCDF_FILE

	Splus
	> iod <- surf(c("u","v","u:w","p","u.csat"))
	> d <- read(iod,"1997 apr 23 00:00","1997 apr 23 01:00")
	> close(iod)

There are two other arguments to read(): dt and avgperiod.

dt is the desired resolution of the data.  If any variables in
the file are higher resolution than dt, then they are averaged
to dt, with the specified average period (see average.ats).

If you do a read of a combination of 5 and 1 minute data, and
specify dt=60:

	iod <- surf(c("u","v","u.prop","v.prop"))
	d <- read(iod,"1997 apr 23 00:00","1997 apr 23 01:00",dt=60)

the resulting matrix will contain 1 minute data, with NA's inserted
into the sonic data - no interpolation, or last-value-repeated
support yet.

If you do:

	iod <- surf(c("u","v","u.prop","v.prop"))
	d <- read(iod,"1997 apr 23 00:00","1997 apr 23 01:00")

then you'll get 5 minute data, with block averages of the 1 minute
data, because the default for dt is 300 secs.


More to come...

  Cocklebur rebooted at around 12:20 GMT this morning.
Marigold came back to life successfully; but, cosmos did
not. 
  Cocklebur went down around 10:25 GMT. It was down for
approximately 2 hours.

1. su -		# become super-user on aster (not cocklebur or platypus)
2. csh		# run the cshell
3. cd /usr/local/adm/backup
4. ls *.log

   The latest version of the backup script creates log files with names like:

	backup.YYMMDD.N.log

   N is the tape number of the last backup.  Currently we cycle through
   6 backup tapes, so N should be 1-6.

5. Put the next-in-sequence backup tape in the drive. 

6. setenv TAPE /dev/rmt/0mn	# drive 0, 8500 density, not compressed

7. ./backup		# remember to use the dot-slash

8. the tape will rewind and eject when the script finishes successfully.
   Check the log file for errors.


The backup script creates an ufsdump file on the tape for each of these
file systems:
	/
	/usr
	/var
	/export
	/opt
	/home
	/pam

Note that /data is not backed up.  Files on that disk should be
backed up with bkfiles.

kansas.net seems to be shutting down our connection, often every
30 minutes.  I Emailed Brian Zimmerman and asked him to look into it.

The connection time is often exactly 30 minutes, so that it does not
appear to be due to a noisy phone line or loss of carrier.

I've put a ping of kansas-net into my crontab to run every 10 minutes.
That should bring the connection back up so that Boulderites
can log in.

Here is the initial configuration of 4 Freewaves before they left
for CASES.



##############################################################
                cocklebur term/0                marigold port/a
                ipdptp0

                560-0563                        560-0575
                master calling 560-0575         slave
Baud:           115200                          38400

FreqKey:        11
Max packet:     7
Min packet:     9
Xmit rate       1
RF data rate:   3
RF Xmit power   5 in lab, 9 in field


##############################################################
                cocklebur term/1                cosmos port/a
                ipdptp1

                560-0512                        560-0573
                master calling 560-0575         slave
Baud:           115200                          38400

FreqKey:        E
Max packet:     7
Min packet:     9
Xmit rate       1
RF data rate:   3
RF Xmit power   0 in lab, 9 in field

			Files of Interest

These files are all on cocklebur, don't look for them on aster!

	/etc/asppp.cf		main configuration file
	/etc/uucp
		Systems		dial-out config: phone #s, accounts,
					login "chat" script
		Devices		dial-out devices
		Dialers		modem AT command sequences

	/etc/log/asppp.log	log file, quite useful

	/usr/local/adm/serialports/ppp_direct.pmadm
				script to configure dial-in terminal
				ports for adam connections

/etc/asppp.cf describes the point-to-point interfaces

	ipdptp0	-> marigold
	ipdptp1	-> cosmos
	ipdptp2 -> kansas-net

The file /etc/uucp/Systems contains further information on dial-out
connections. The phone number and account name for the kansas-net
internet service are listed in this file.  In addition, the entry
for kansas-net contains the expect/send strings that are used 
during the login "chat" session.  The kansas-net entry specifies
ACU as the call device.  The description of the ACU is found in
/etc/uucp/Devices. The last field in the entry for ACU is "usrv32-ec"
which is an entry for a US Robotics modem in /etc/uucp/Dialers.
The "usrv32-ec" entry in Dialers is where the modem AT command
sequence is specified.


	Diagnosing the PPP connection to kansas.net.  

  The results of the dial sequence and the login chat are logged in asppp.log.
A good login should look like the following:

00:08:00 process_ipd_msg: ipdptp2 needs connection
conn(kansas-net)
Trying entry from '/etc/uucp/Systems' - device type ACU.
Device Type ACU wanted
Trying device entry 'cua/2' from '/etc/uucp/Devices'.
processdev: calling setdevcfg(ppp, ACU)
gdial(usrv32-ec) called
Trying caller script 'usrv32-ec' from '/etc/uucp/Dialers'.
expect: ("")
got it
sendthem (DELAY
APAUSE
TE1V1X1Q0S2=255S12=255&A3&K1&H1&R2&I0&M4&B1&N0&W^M)
expect: (OK^M)
ATE1V1X1Q0S2=255S12=255&A3&K1&H1&R2&I0&M4&B1&N0&W^M^M^JOK^Mgot it
sendthem (ECHO CHECK ON
A^JATTDDTT1133116677882233880022^M^M)
expect: (CONNECT)
^M^JCONNECTgot it
STTY crtscts,crtsxoff
getto ret 15
expect: ("")
got it
expect: ("")
got it
sendthem (^M^M)
expect: (ogin:)
lcome to Wheat State Internet Se.Wk+^K^I^IuIQ^AE%5)5)5)5)1=^]%9i^A^?^M^Jlogin:go
t it
sendthem (sssf^M)
expect: (ssword:)
 sssf^M^JPassword:got it
sendthem XXXX--DELETED FROM LOG
expect: ("")
got it
STTY crtscts,crtsxoff
call cleanup(0)^M
00:08:48 start_ip: IP up on interface ipdptp2, timeout set for 3600 seconds 



		Diagnosing PPP RF connections

Log file:  cocklebur:/etc/log/asppp.log

The asppp.log file is useful to watch. You may want to leave a term window
on cocklebur doing a "tail -f asppp.log".  

Every time the carrier line drops on a Freewave modem, even for a second,
you will see entries like so:

01:53:06 process_ipd_msg: interface ipdptp1 has disconnected
01:53:06 disconnect: disconnected connection from  ipdptp1
01:53:12 start_ip: IP up on interface ipdptp1, timeout set for 600 seconds

This indicates that carrier was lost momentarily on the Freewave
connection to cosmos.  It was back up in 6 seconds.  The socket
connections to the ADAM stay up, the adam keeps running and no data
is lost during these short interruptions.  I saw these carrier drops
every few minutes when testing the freewaves in the lab, with the
antennas 2 feet apart, on a low power setting.


If the RF modems are powered up with carrier, then the aspppls process
should be active on the serial ports.  
To check for aspppls enter:

	ps -t term/0
	ps -t term/1

These commands should show the aspppls process on the port.
   PID TTY      TIME CMD
  7877 term/0   0:00 aspppls

If the carrier detect on a serial port is too active, going up and
down repeatedly, then after a while the Sun will start to ignore it,
namely the port monitor will not spawn aspppls when the carrier detect line
is raised..  If aspppls is not active on the port, then the adam will
not be able to establish a ppp connection.

There is a crontab entry on the root account on cocklebur, which
checks every 10 minutes that aspppls is active on the two ports,
and if not, runs /usr/local/adm/serialports/ppp_direct.pmadm which
restarts the port monitor.  You must be root to restart the port monitor.

	Status of Adam connection

To view statistics of the Adam data connection, enter:

	sshow cosmos 0
or 	sshow marigold 0

This will display the data from the status channel 0 from the adam
as follows:

c:000 cosmos mem:50560 nent:100 nlost:0 nlate:0 pS:0 pD:0 dT:5100 msec, blocked:
863, lost:0, writerr:0 buf:80624

"lost:" shows the number of data samples that have been lost due
to network stoppages.  The number of severe write errors on the 
data socket are also shown under writerr.   "buf:" shows the available
network buffer size in bytes.  As the network becomes clogged the
available buffer size will shrink to 0.


        Viewing Setup and RF Statistics of a Freewave Modem


Disconnect the serial cable between cocklebur and the freewave.
This will cause a carrier detect drop on the line and processes
that are using the serial port will exit.  Then do:

        kermit -l /dev/cua/N -b 19200 -c

        N is 0 for the marigold freewave, 1 for cosmos

If you get a "? can't open device" wait a few moments and try again.

Once kermit says it is connected, reconnect the cable.  Press the reset
button on the front of the modem, and use the menus to configure
the modem and display the RF statistics.

	INTERNET 
	  |
	potwinpm.kansas.net
	206.103.127.129
	  |
	  |
	  | ppp over phone modem
	  |
	  |
----------------------------------------------------------------------------
|	term/2			COCKLEBUR				    |
|     206.103.127.130							    |
|     sssf.kansas.net							    |
|									    |
|									    |
|     cocklebur			adam-ppp			adam-ppp    |
|  128.117.87.28		128.117.87.29			128.117.87.29
|  le0				term/0				term/1      |
-----------------------------------------------------------------------------
  |				|				|
  |				|     PPP over RF modem		|
  | coax ethernet		|				|
  |				|				|
  |				|				|
  |   aster			marigold			cosmos
  |-- 128.117.87.40		128.117.87.32			128.117.87.30
  |				|				|
  |				| vxworks			| vxworks
  |				|				|
  |   platypus			arnica				goldenrod
  |-- 128.117.87.41		128.117.87.9			128.117.87.2
  |
  |
  |   dj
  |-- 128.117.87.42



Note that cocklebur is our big-time network gateway, and has
3 different net names (cocklebur, sssf.kansas.net and adam-ppp)
and 3 IP addresses.  When you're accessing it from outside, address
it as sssf.kansas.net.  When accessing it from the other PAM base
systems, call it cocklebur.


For ftp, telnet, rlogin and Web access, sssf.kansas.net is
the only system that is visible to the outside world.  Therefore, if
you want to ftp to the PAM base from Boulder, you must ftp
to sssf.kansas.net.  Likewise, if you want to ftp from the base to
anywhere on the internet, you must do it from sssf.kansas.net (cocklebur).

Email has been configured so that you can send email from any system
at the base (aster, platypus, cocklebur) to the Internet.
Your Email will have a return address of user@sssf.kansas.net.
It is should be much easier to send email directly from the base,
rather than logging into stout.

Once you're in the field (and the base system is setup) you can have
your email forwarded to Kansas. To do this, log into stout and create
a .forward  file on your home directory containing one line:
	youruserid@sssf.kansas.net


ADAM network:

You can rlogin or rserial the adams from cocklebur, aster or platypus.

Configuration of the PPP service with Wheatstate Telephone

Our nodename: 	sssf.kansas.net
IP address:	206.103.127.130  aka pm1-potwin-130.kansas.net
DNS servers:	primary 206.103.126.60  = tuttle.kansas.net
		secondary: 199.79.146.61 = bbs.tfs.net

Access phone #:	316 782 3802
Account:	sssf


IP address of PPP server at wheatstate:
		206.103.127.129 potwinpm.kansas.net

Contact:	Brian Zimmerman
		Internet Tech Support
		Wheat State Telephone Inc

		http://www.wheatstate.com

		800 442-6835 ext 116
		800 227-4547 - most often at this number