Computing

Restarting the Live Pages:

Login to newsmerd as user observer
Execute the command sudo /etc/init.d/sql_monitors restart
It will ask for a password, put the standard observer password in. This should restart the live page updating.

For Ceduna the instructions are the same, the only difference is that you need to login to sille (131.217.61.174)

Of course, you may not need to do any of this if the only problem is that your browser's cache is in a knot. So If the page refreshes OK on one computer's browser, but not another, then the solution is to clear that computer's browser cache. To do this on newsmerd, click edit, preferences, privacy, cache, clear cache now.

Restarting DAS:

You may need to do this if particularly if the MB correlator bandpass looks like rubbish in one channel. This has recently been found to be neccesary when using profile cor4000.pro.

Exit the DAS program, Q.
After you see the message "Tousches", exit the DAS window
Power cycle the DAS at the rear left hand power switch
After a while, open a new DAS window on the Windows DAS PC.
Load the profile you want.
If it displays a configuration screen showing a baud rate of something other than -27136 then enter 5 to change the baud rate and then enter -27136 as the baud rate. Yes this is weird, but it is what works.

Network timeout

If the fs monitor on newsmerd has an alarm, first try pinging Rakbus and sam26m from newsmerd. If rakbus can not be pinged it may be necessary to restart the rakbus pc, and then issue the command

calu -m sam26m -r rakbus

in a terminal on newsmerd. You will then have to restart STALM on hobart and the pcfs monitor on newsmerd. You may first need to exit the fs monitor on newsmerd using CTRL-C. Then check that the ACT antenna monitor is updating correctly on newsmerd. (Brett thinks these last 2 sentences only apply if you are doing VLBI. single dish users need not worry.)

Correlator boot floppy:

The multibeam correlator at Hobart has a Linux boot floppy which makes a bootp request for network booting. If this should become corrupted a new copy can be created by :

Copy the file smc_image.flo from /home/oper/correlator on Hobart, to a Linux machine with a 3.5 inch floppy disk
Issue the command dd if=smc_image.flo of=/dev/fd0 You may need to be root to do this. It should write 28 records in/out.

Copy mark5 data to a file:

First mount a drive to copy to. As root on mkv, do the following commands

mkdir /mnt/hovsi_removable

mount hovsi.phys.utas.edu.au:/data/removable /mnt/hovsi_removable

then do: mkdir /mnt/hovsi_removable/mk5disk2file

cd /mnt/hovsi_removable/mk5disk2file

and you have a place to store the linux files you will create.

As oper on mkv terminal screen (not the console), and with the module containing desired data selected on the recorder, and with Mark5A software already running (the state the recorder is normally in) do the following:

tstMark5A the prompt will come back ">"

disk2file=/mnt/hovsi_removable/mk5disk2file/vt05ahotest:0:1000000000:w;

The last command will write a 1 GB file from record pointer position zero and call the file vt05ahotest. You should substitute the file name, start byte and end bytes you want. refer to Mark5A command set for details of using disk2file.

During the disk2file you can do "disk2file?" to find out where it is up to, eg:

> disk2file?

disk2file? 0 : active : /mnt/hovsi_removable/mk5disk2file/vt05ahotest : 0 : 899481600 : 1000000000 : w ;

> disk2file?

disk2file? 0 : inactive : /mnt/hovsi_removable/mk5disk2file/vt05ahotest ; >

In the above examples the process has completed 899 MB the first time and was completed at the second query.