On 13 Nov 2011, at 02:57, Petr Tsatsin wrote:
Hello,
I'm trying to restart my simulation on queenbee using simfactory. I tried "./sim submit --from-restart-id ..." and I got
What was the full command that you used?
You should not normally need the --from-restart-id option - SimFactory uses this internally. It is supposed to work out which restart to recover from automatically.
You should be able to do
sim submit <simulation>
and it should queue a restart which recovers. If this doesn't work, can you post your log.txt file (from the simulation directory) so we can see what went wrong?
"Simulation Factory: [log] Submitting: [log] Using restart id "0002" [log] Requested recovering from restart id "0" [log] Could not find checkpoint files in restart id "0000" Aborting.'' Inside my parameter file for this simulation I set up HDF5 check pointing in following way: "IO::out_dir = $parfile IOHDF5::checkpoint = "yes" IO::checkpoint_every = 100000 IO::checkpoint_file = $parfile IO::checkpoint_dir = $parfile IO::recover="auto" IO::recover_file = $parfile IO::recover_dir = $parfile" I checked that h5 check point files are actually inside the output-0000 directory. Any suggestions how can I restart the my simulation from HDF5 checkpoint with a simfactory? Thank you. -- Petr Tsatsin Graduate Student- Department of Physics Charles E. Schmidt College of Science Florida Atlantic University (561)-297-3386 ptsatin@fau.edu, ptsatsin@gmail.com
Users mailing list Users@einsteintoolkit.org http://lists.einsteintoolkit.org/mailman/listinfo/users
Petr
In most cases, a simple "submit" (with the same simulation name) should suffice. --from-restart-id allows restarting from a particular restart, in case you don't want to restart from the previous restart.
Simfactory has a parameter file simfactory/etc/parfiles/recover.par. Could you try this? You would submit it on a single process, let it run for a few minutes, and then submit it again:
./bin/sim submit recover --parfile=simfactory/etc/parfiles/recover.par --walltime=0:10:0
and later
./bin/sim submit recover --parfile=simfactory/etc/parfiles/recover.par --walltime=0:10:0 This should work.
-erik
On Sat, Nov 12, 2011 at 8:57 PM, Petr Tsatsin ptsatsin@fau.edu wrote:
Hello,
I'm trying to restart my simulation on queenbee using simfactory. I tried "./sim submit --from-restart-id ..." and I got "Simulation Factory: [log] Submitting: [log] Using restart id "0002" [log] Requested recovering from restart id "0" [log] Could not find checkpoint files in restart id "0000" Aborting.'' Inside my parameter file for this simulation I set up HDF5 check pointing in following way: "IO::out_dir = $parfile IOHDF5::checkpoint = "yes" IO::checkpoint_every = 100000 IO::checkpoint_file = $parfile IO::checkpoint_dir = $parfile IO::recover="auto" IO::recover_file = $parfile IO::recover_dir = $parfile" I checked that h5 check point files are actually inside the output-0000 directory. Any suggestions how can I restart the my simulation from HDF5 checkpoint with a simfactory? Thank you. -- Petr Tsatsin Graduate Student- Department of Physics Charles E. Schmidt College of Science Florida Atlantic University (561)-297-3386 ptsatin@fau.edu, ptsatsin@gmail.com
Users mailing list Users@einsteintoolkit.org http://lists.einsteintoolkit.org/mailman/listinfo/users
Hello Erik, I haven't found any recover.par. May be I have an older version of the simfactory. So, I tried just to run with my par file. First i used this command ./simfactory/sim create-submit recover_test2 --parfile=par/myparfile.par --procs=8 --walltime=0:10:0 then ./simfactory/sim submit recover_test2 --parfile=par/myparfile.par --procs=8 --walltime=0:10:0 and I got the following error message: Simulation Factory: [log] Submitting: [log] Restart id "0000" is active Aborting.
Thank you. Petr.
Here is a LOG file: LOG FILE for simulation "recover_test2" ================================================================================
-------------------------------------------------------------------------------- 2011-11-21 10:08:19 ettest04@qb4.loni.org: Skeleton created Job directory: "/scratch/ettest04/simulations/recover_test2" Machine: "queenbee" Simulation id: "simulation-recover_test2-queenbee-qb4.loni.org-ettest04-2011.11.21-10.08.19-8396" Source dir: "/home/ettest04/Cactus" Configuration: "sim" Config id: "config-sim-qb4.loni.org-home-ettest04-Cactus" Build id: "build-sim-qb4.loni.org-ettest04-2011.11.08-19.43.33-31159" Executable: "/home/ettest04/Cactus/exe/cactus_sim" Option list: "configs/sim/OptionList" Script file: "/home/ettest04/Cactus/configs/sim/ScriptFile" Parameter file: "par/rns_64_25km_1lev_tvd_vanleerMC2_hlle_MADM_1.432_HarmF_2_HarmN_1_g12.par" --------------------------------------------------------------------------------
-------------------------------------------------------------------------------- 2011-11-21 10:08:19 ettest04@qb4.loni.org: Submitting: Using restart id "0000" Not recovering since there are no restart ids Created restart directory Activated restart directory Created job script Created parameter file About to submit job Executing: cd /scratch/ettest04/simulations/recover_test2/output-0000 && { qsub SIMFACTORY/ScriptFile ; } Submitted job Job id is "583945" --------------------------------------------------------------------------------
-------------------------------------------------------------------------------- 2011-11-21 10:47:32 ettest04@qb4.loni.org: Submitting: Restart id "0000" is active --------------------------------------------------------------------------------
Tha
On Sun, Nov 13, 2011 at 8:48 AM, Erik Schnetter schnetter@cct.lsu.eduwrote:
Petr
In most cases, a simple "submit" (with the same simulation name) should suffice. --from-restart-id allows restarting from a particular restart, in case you don't want to restart from the previous restart.
Simfactory has a parameter file simfactory/etc/parfiles/recover.par. Could you try this? You would submit it on a single process, let it run for a few minutes, and then submit it again:
./bin/sim submit recover --parfile=simfactory/etc/parfiles/recover.par --walltime=0:10:0
and later
./bin/sim submit recover --parfile=simfactory/etc/parfiles/recover.par --walltime=0:10:0 This should work.
-erik
On Sat, Nov 12, 2011 at 8:57 PM, Petr Tsatsin ptsatsin@fau.edu wrote:
Hello,
I'm trying to restart my simulation on queenbee using simfactory. I tried "./sim submit --from-restart-id ..." and I got "Simulation Factory: [log] Submitting: [log] Using restart id "0002" [log] Requested recovering from restart id "0" [log] Could not find checkpoint files in restart id "0000" Aborting.'' Inside my parameter file for this simulation I set up HDF5 check
pointing in
following way: "IO::out_dir = $parfile IOHDF5::checkpoint = "yes" IO::checkpoint_every = 100000 IO::checkpoint_file = $parfile IO::checkpoint_dir = $parfile IO::recover="auto" IO::recover_file = $parfile IO::recover_dir = $parfile" I checked that h5 check point files are actually inside the output-0000 directory. Any suggestions how can I restart the my simulation from HDF5 checkpoint with a simfactory? Thank you. -- Petr Tsatsin Graduate Student- Department of Physics Charles E. Schmidt College of Science Florida Atlantic University (561)-297-3386 ptsatin@fau.edu, ptsatsin@gmail.com
Users mailing list Users@einsteintoolkit.org http://lists.einsteintoolkit.org/mailman/listinfo/users
-- Erik Schnetter schnetter@cct.lsu.edu http://www.cct.lsu.edu/~eschnett/
On 21 Nov 2011, at 18:20, Tsatsin Petr wrote:
Hello Erik, I haven't found any recover.par. May be I have an older version of the simfactory.
If you are using an old version of simfactory, this is important information. Please could you post the output of
cd simfactory svn info svn status
so that we can see what version you are using?
So, I tried just to run with my par file. First i used this command ./simfactory/sim create-submit recover_test2 --parfile=par/myparfile.par --procs=8 --walltime=0:10:0 then ./simfactory/sim submit recover_test2 --parfile=par/myparfile.par --procs=8 --walltime=0:10:0
You should not need the "--parfile=par/myparfile.par --procs=8 --walltime=0:10:0" options when you submit the recovery job because SimFactory will automatically determine these based on the first job. Having said that, they should not cause any problem.
and I got the following error message: Simulation Factory: [log] Submitting: [log] Restart id "0000" is active Aborting.
That should not happen. Are you sure you are using SimFactory 2? We will see from the output of svn info that you send.
Here it is.
svn info: [ettest04@qb4 simfactory]$ svn info Path: . URL: https://svn.cct.lsu.edu/repos/numrel/simfactory/branches/ET_2011_05 Repository UUID: 335acd41-667a-45b5-96c6-ccfbb2e2c6d2 Revision: 1473 Node Kind: directory Schedule: normal Last Changed Author: hinder Last Changed Rev: 1472 Last Changed Date: 2011-10-04 07:51:58 -0500 (Tue, 04 Oct 2011)
[ettest04@qb4 simfactory]$ svn status ? udb.pm
I'm not sure that I'm using a last version. This is might be a solution. Thank you. Petr.
On Mon, Nov 21, 2011 at 1:37 PM, Ian Hinder ian.hinder@aei.mpg.de wrote:
On 21 Nov 2011, at 18:20, Tsatsin Petr wrote:
Hello Erik, I haven't found any recover.par. May be I have an older version of the simfactory.
If you are using an old version of simfactory, this is important information. Please could you post the output of
cd simfactory svn info svn status
so that we can see what version you are using?
So, I tried just to run with my par file. First i used this command ./simfactory/sim create-submit recover_test2 --parfile=par/myparfile.par --procs=8 --walltime=0:10:0 then ./simfactory/sim submit recover_test2 --parfile=par/myparfile.par --procs=8 --walltime=0:10:0
You should not need the "--parfile=par/myparfile.par --procs=8 --walltime=0:10:0" options when you submit the recovery job because SimFactory will automatically determine these based on the first job. Having said that, they should not cause any problem.
and I got the following error message: Simulation Factory: [log] Submitting: [log] Restart id "0000" is active Aborting.
That should not happen. Are you sure you are using SimFactory 2? We will see from the output of svn info that you send.
-- Ian Hinder http://numrel.aei.mpg.de/people/hinder
Petr
No, you are not using the latest version. Since there are substantial changes to your version (which should still work), you can try to update. https://docs.einsteintoolkit.org/et-docs/Tutorial_for_New_Users has instructions for checking out a new version of the Einstein Toolkit; Simfactory is at https://svn.cct.lsu.edu/repos/numrel/simfactory2/trunk.
-erik
On Mon, Nov 21, 2011 at 1:55 PM, Tsatsin Petr ptsatsin@gmail.com wrote:
Here it is.
svn info: [ettest04@qb4 simfactory]$ svn info Path: . URL: https://svn.cct.lsu.edu/repos/numrel/simfactory/branches/ET_2011_05 Repository UUID: 335acd41-667a-45b5-96c6-ccfbb2e2c6d2 Revision: 1473 Node Kind: directory Schedule: normal Last Changed Author: hinder Last Changed Rev: 1472 Last Changed Date: 2011-10-04 07:51:58 -0500 (Tue, 04 Oct 2011) [ettest04@qb4 simfactory]$ svn status ? udb.pm I'm not sure that I'm using a last version. This is might be a solution. Thank you. Petr. On Mon, Nov 21, 2011 at 1:37 PM, Ian Hinder ian.hinder@aei.mpg.de wrote:
On 21 Nov 2011, at 18:20, Tsatsin Petr wrote:
Hello Erik, I haven't found any recover.par. May be I have an older version of the simfactory.
If you are using an old version of simfactory, this is important information. Please could you post the output of cd simfactory svn info svn status so that we can see what version you are using?
So, I tried just to run with my par file. First i used this command ./simfactory/sim create-submit recover_test2 --parfile=par/myparfile.par --procs=8 --walltime=0:10:0 then ./simfactory/sim submit recover_test2 --parfile=par/myparfile.par --procs=8 --walltime=0:10:0
You should not need the "--parfile=par/myparfile.par --procs=8 --walltime=0:10:0" options when you submit the recovery job because SimFactory will automatically determine these based on the first job. Having said that, they should not cause any problem.
and I got the following error message: Simulation Factory: [log] Submitting: [log] Restart id "0000" is active Aborting.
That should not happen. Are you sure you are using SimFactory 2? We will see from the output of svn info that you send. -- Ian Hinder http://numrel.aei.mpg.de/people/hinder
-- Petr Tsatsin Graduate Student- Department of Physics Charles E. Schmidt College of Science Florida Atlantic University (561)-297-3386 ptsatin@fau.edu, ptsatsin@gmail.com
users@lists.einsteintoolkit.org