The "HORSEFLY" Run
------------------
Feb 15-18 2008
VNC
there's a session running on flytrap that you can tunnel out to your local
machine like:
xterm1:
ssh -L 5911:localhost:5911 [email protected]
ssh -L 5911:localhost:5901 casper@flytrap
xterm2:
vncviewer -shared localhost:1
pwd starts with 4 caps
MAKE PULSAR POLYCO
on gasp0 machine at GB: makepoly 0329+54 1430 1 s 12
[ask joeri]
MAKE EPHEM:
(got rajd decd from atnf db)
ataephem --radec 53.24737,54.57877 --stdout > PSRB0329+54.ephem [wrong!]
0329+54 doesn't show up, as ataephem takes fracRA not RA in degrees, as above
-> remade with fracRA not RADeg [possibly still a little off], 0329 now detected.
HORSESHOE POINTING:
changed horseshoe.csh (save old one as horseshoe-orig.csh) as it was counting
some X and Y for two independent _sky pointings_.
cat Horseshoe.csh | awk '{print $2}' | sort | perl -pi -e 's/\n/,/'
-> outputs antenna list for
atatrackephem 1A,1C,1D,1E,1F,1G,1J,1K,2A,2B,2C,2F,2G,2H,2H,2J,2K,2L,3C,3D,3E,3G,4F,4J,4K,4L,5B,5C,5E,5H PSRB0329+54.ephem
PULSAR TEST COMMANDS
file="/media/disk1/2008_02_17/PSRB0329+54-17.log -r 300"
ln -sf /usr/src/sigproc-3.7/fold .
time filterbank-ata-03 $file
ls -1 *fil | awk '{print "dedisperse "$1" -d 26.776000 > "$1".tim; echo "$1}' | csh -f
ls -1 *tim | awk '{print "./fold "$1" -p ./polyco.dat.0329+54-54514.8-s -stream -n 128 -m 2 > "$1".fold"}' | csh -f
(cat *.fold | grep -v DONE ; echo '#DONE') | pgplotter 4 11 '/vps'
(cat *.fold | grep -v DONE ; echo '#DONE') | pgplotter 4 11
CATALOG
made a bright-psrs.txt file and added them to ephemeris DB using
ataupdatecatalog psr PSRB0950+08 9.885919 7.92660
etc
OBSERVING
cd /home/casper/obsscript
emacs -nw launch.sh
to change savedir
nohup ./launch.sh &
*DISK=/media/disk1/2008_02_15
>pulsar runs done 'by hand' a la
./gulp -i eth1 > file & &
sleep 900; kill 1234
psr032954.log - psr032954-5.log
*DISK=/media/JOERI_FLYEYE_2/2008_02_15
1st horseshoe data_54512.24 - data_54513.00
PSRBpsr032954-6.log - PSRBpsr032954-8.log
*DISK=/media/JOERI_FLYEYE_1/2008_02_16
2nd horshoe
seems to have failed about 5 min (~3GB in) at each of the 3 one-hour runs.
testing + fix below, had to reload USB kernel module
--next crab
ataradecephem 03:32:59.3758 54-34-43.518 now +1hours > B0329+54-4.ephem
atatrackephem 1A,1C,1D,1E,1F,1G,1J,1K,2A,2B,2C,2F,2G,2H,2H,2J,2K,2L,3C,3D,3E,3G,4F,4J,4K,4L,5B,5C,5E,5H B0329+54-4.ephem
PSRB032954-12.log [2 minutes, 60% ringbuffer!]
[unmounted J_BKY_2+3, disk-1]
PSRB032954-13.log 11.21pm -- 11.25 ring buffer overflow
>DISK _1 now extremely slow]
[unmounted _1, mounted _3]
[umounted everything; modprobe -r usb_storage; modprobe usb_storage; all
devices rescanned automatically; mounted only _1; appears better]
5 minute run on 0950: only 3% ringbuffer use
/media/JOERI_FLYEYE_1/2008_02_16/PSRB0950+08-01.log
5 minute on 0329:
/media/JOERI_FLYEYE_1/2008_02_16/PSRB0329+54-15.log
--back to horseshoe
at 12:37pm, to disk J_F_1
--1933+16
/media/JOERI_FLYEYE_1/2008_02_16/PSRB1933+16-01.log [15 minutes]
>testing rsync while observing to free up some disk space
rsync -avP --bwlimit=5000 adds 0.5-0.6 to load
gulp+rsync-5M load=>2.0; ring buffer 40% after 2min
gulp+rsync-4M load=1.5; ring buffer 30% after 2min (so too much)
gulp+rsync-3M load=1.5; ring buffer 2% after 2min (i.e. good)
-- give this one a try
--back to horsehoe, with rsync at 3MBps (rsync -avP --bwlimit=5000)
doing fine at long-termload=1.50-1.60
very little packet loss over 12-hr run
>at some point during this run the 1x,2x,3x telescope might have stopped
responding. only noticed someting strange when trying to repoint later.
will write directly to archive to avert disk crunch. this data is not the best anyway.
*DISK=/lordoftheflies/dataarchive/20080216
[i.e. writing directly to archive]
[will 'gzip --fast' data later to remove zero-channels
turns out to compress files in real time from 36 to 8GB]
>rick power cycled broken antennae
restarting around 3pm, MJD around 54514.96
atasetskyfreq a 1430
ataephem -v PSRB0329+54
*DISK=/media/disk1/2008_02_17/
PSRB0329+54-16.log [5min]
PSRB0329+54-17.log [10min][psr visible in pretty much all dishes]
-turned on walshing
PSRB0329+54-18-WALSHING_ON.log [15min]
PSRB0531+21-01-WALSHING_ON.log [30min]
-turned off walshing
PSRB0531+21-02-WALSHING_OFF.log [30min]
-northern horseshoe for rest of night
/ataarchive/2008/02/15/Horseshoe-Changed.csh
last file MJD=54515.22
-changed to southern horseshoe with
/ataarchive/2008/02/15/SouthPointing-Changed.csh
[edited from Geoff's email to remove dishes that were used twice]
*DISK=/media/JOERI_FLYEYE_3
files /media/JOERI_FLYEYE_3/2008_02_18 ; MJD 54515.27-54515.93
> the 1x antennae were 'stopped' (externally I presume) at some point it appears
> the 2x antennae were 'stopped' around 3pm.
- at 4:50pm power cycled flyeye alpha by typo in the ethernet power strip.
- you can't both use lof local FW and copy to it from flyeye -- slows down both.
COPY TO ARCHIVE:
/media/JOERI_FLYEYE_1 Done [plugged in LOF as /dev/disk9s1] [colby 2 berkeley?]
/media/JOERI_FLYEYE_2 Done
/media/JOERI_FLYEYE_3 Done
/media/disk1 Done
Gzipping 11-IF data Done [/Volumes/jack/dataarchive/20080216/]
HOW TO MOUNT FIREWIRE ON G4
http://casper.berkeley.edu/wiki/index.php?title=FlyEye_Node