Parallel rerun in cluster

Andy_bm · November 24, 2011, 11:13

Dear Foamers.
I have same problem with rerun my case after cluster turn off and again turn on.For example, in first day I put my case in queue calculation, I use the following script:

Code:

#PBS -l walltime=24:00:00,nodes=4:ppn=8
cd ~/OpenFOAM/Andrey-2.0.1/run/SWkw
mpirun -np 32 -machinefile $PBS_NODEFILE rhoCentralFoam -parallel | tee -a log.std

In the end of the working day cluster turn off. On the next day cluster turn on, but my case don't calculation in cluster.There are two variants:
1) Case is in queue but don't calculation.Use time in qstat = constant 00:00:00.
2) Case is not in queue, and in working directory I have file .e , .o created at the moment of switching-off cluster.

in controlDict starttime set to latestTime.

What probably to make,after cluster turn on,that openfoam continued to calculation?

Thanks.

akidess · November 25, 2011, 02:40

Do you have superuser rights on the cluster?

Andy_bm · November 25, 2011, 13:38

No, I have not root rights,but I can say superuser on this cluster that it is possible to make.

akidess · November 27, 2011, 06:27

First, you should ask the administrator to not turn off the cluster so that your jobs can finish over night. If that's absolutely not possible, you'll have to ask him to set up an init.d script to restart your jobs on reboot.

Andy_bm · November 27, 2011, 07:16

Unfortunately,not turn off the cluster is impossible.
Could you tell as to adjust this init.d script?

November 24, 2011, 11:13	Parallel rerun in cluster	#1
Andy_bm Member Join Date: Nov 2011 Posts: 32 Rep Power: 13	Dear Foamers. I have same problem with rerun my case after cluster turn off and again turn on.For example, in first day I put my case in queue calculation, I use the following script: Code: #PBS -l walltime=24:00:00,nodes=4:ppn=8 cd ~/OpenFOAM/Andrey-2.0.1/run/SWkw mpirun -np 32 -machinefile $PBS_NODEFILE rhoCentralFoam -parallel \| tee -a log.std In the end of the working day cluster turn off. On the next day cluster turn on, but my case don't calculation in cluster.There are two variants: 1) Case is in queue but don't calculation.Use time in qstat = constant 00:00:00. 2) Case is not in queue, and in working directory I have file .e , .o created at the moment of switching-off cluster. in controlDict starttime set to latestTime. What probably to make,after cluster turn on,that openfoam continued to calculation? Thanks.

Thread Tools	Search this Thread
Show Printable Version Email this Page	Search this Thread: Advanced Search
Display Modes
Linear Mode Switch to Hybrid Mode Switch to Threaded Mode

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Script to Run Parallel Jobs in Rocks Cluster	asaha	OpenFOAM Running, Solving & CFD	12	July 4, 2012 22:51
Parallel Solving with HPC Cluster?	bah	OpenFOAM Running, Solving & CFD	8	October 13, 2011 03:08
Parallel cluster solving with OpenFoam? P2P Cluster?	hornig	OpenFOAM Programming & Development	8	December 5, 2010 16:06
Unable to run OF in parallel on a multiple-node cluster	quartzian	OpenFOAM	3	November 24, 2009 13:37
Help: how to realize UDF on parallel cluster?	Haoyin	FLUENT	1	August 6, 2007 13:53

November 25, 2011, 02:40		#2
akidess Senior Member Anton Kidess Join Date: May 2009 Location: Germany Posts: 1,377 Rep Power: 28	Do you have superuser rights on the cluster?

November 25, 2011, 13:38		#3
Andy_bm Member Join Date: Nov 2011 Posts: 32 Rep Power: 13	No, I have not root rights,but I can say superuser on this cluster that it is possible to make.

November 27, 2011, 06:27		#4
akidess Senior Member Anton Kidess Join Date: May 2009 Location: Germany Posts: 1,377 Rep Power: 28	First, you should ask the administrator to not turn off the cluster so that your jobs can finish over night. If that's absolutely not possible, you'll have to ask him to set up an init.d script to restart your jobs on reboot.

November 27, 2011, 07:16		#5
Andy_bm Member Join Date: Nov 2011 Posts: 32 Rep Power: 13	Unfortunately,not turn off the cluster is impossible. Could you tell as to adjust this init.d script?