Logging in:
To log into the music cluster you must first log into one of the non-music SSLI computers, and then log into music from there. Please do not do any computing using lacha or any SSLI computer not named musicX.
The following should get you on to music1:
% ssh login@lacha.ee.washington.edu
% ssh login@music1
pmake:
To view a lost of all the jobs that are currently running use
cctrl -jobs -all:
(note: path for cctrl: /usr/local/sbin/cctrl)
music1 [508] cctrl -jobs -all
timeout = 2.500000 secs, 2 retries
Getting job list from all
pmake: Lockfile owned by you -- ignoring it
HOST USER JOB PID PORT FROM START TIME COMMAND
music3 bartels 917182588 13868 1001 music1 15:03 0:05 sh
music3 bartels 917186470 13646 1001 music1 15:03 0:05 sh
music3 bartels 917182451 14706 1001 music1 15:03 0:05 sh
music3 bartels 917230764 15480 1001 music1 15:03 0:05 sh
To check the status of the computers on the music cluster use
reginfo:
music1 [501] reginfo
HOST ARCH STATUS INDEX CLIENTS
music1 4 M available 256 ALL
music2 4 available 256 ALL
music3 4 available 256 ALL
music4 4 available 256 ALL
music5 4 available 256 ALL
music6 4 available 256 ALL
music7 4 available 256 ALL
music8 4 available 256 ALL
music9 4 available 256 ALL
Network 43: 9 host(s) total, 9 available unrestricted, 0 restricted, 0 down
A third command is
rexport. This can send a single command line or file with a list of command lines to the cluster.
music1 [521] rexport echo hi
*** exported to music3 (id 917195342)
hi
music1 [522] rexport -f list_of_commands
*** exported to music3 (id 917184719)
hi 1
*** exported to music3 (id 917189587)
hi 2
*** exported to music3 (id 917207473)
hi 3
When setting up a header file for
emtrain_parallel.JHU_ws06 set the following two variables as follows:
RUN_EMACC="rexport -J 36 -f "
RUN_EMTRAIN"rexport -J 36 -f "
When setting up a header file for
viterbi_parallel.JHU_ws06 set the following variable as follows:
RUN_DISTRIBUTED="rexport -J 99 -f "