repeating postprocessing job – in #9: CCLM

in #9: CCLM

<p> Hello, </p> <p> It appears that I have finished an experiment using the starter and stored all the variables I need. Some of the time series are missing however. Now I would like to repeat the post-processing in order to create additional time series of some missing values. <br/> Is that possible? How this may be done? <br/> Regards <br/> Simon </p>

  @redc_migration in #53dcddc

<p> Hello, </p> <p> It appears that I have finished an experiment using the starter and stored all the variables I need. Some of the time series are missing however. Now I would like to repeat the post-processing in order to create additional time series of some missing values. <br/> Is that possible? How this may be done? <br/> Regards <br/> Simon </p>

repeating postprocessing job

re wrote

Hello,

It appears that I have finished an experiment using the starter and stored all the variables I need. Some of the time series are missing however. Now I would like to repeat the post-processing in order to create additional time series of some missing values.
Is that possible? How this may be done?
Regards
Simon

View in channel
<p> Hi, <br/> you can have a look into your jobs directory of your experiment and rerun the post(yourEXID).job. You can either run the whole postprocessing job or copy the desired lines out of this script into your terminal. <br/> Hope this helps. <br/> Best regards, <br/> Merja </p>

  @merjatölle in #f521806

<p> Hi, <br/> you can have a look into your jobs directory of your experiment and rerun the post(yourEXID).job. You can either run the whole postprocessing job or copy the desired lines out of this script into your terminal. <br/> Hope this helps. <br/> Best regards, <br/> Merja </p>
MT replied

Hi,
you can have a look into your jobs directory of your experiment and rerun the post(yourEXID).job. You can either run the whole postprocessing job or copy the desired lines out of this script into your terminal.
Hope this helps.
Best regards,
Merja

View in channel
<p> Dear Merja, <br/> My question was not clear enough probably. <br/> After finishing a several year run I already have no cclm monthly output files, but only those in directories …/chain/arch and /work/xxx/restarts . <br/> Some of the time series for all months are missing. Now I would like to repeat the post-processing in order to create additional time series of some missing values for each month of my period. Can this be done? <br/> Many thanks for your help <br/> Simon </p>

  @redc_migration in #486a424

<p> Dear Merja, <br/> My question was not clear enough probably. <br/> After finishing a several year run I already have no cclm monthly output files, but only those in directories …/chain/arch and /work/xxx/restarts . <br/> Some of the time series for all months are missing. Now I would like to repeat the post-processing in order to create additional time series of some missing values for each month of my period. Can this be done? <br/> Many thanks for your help <br/> Simon </p>
re replied

Dear Merja,
My question was not clear enough probably.
After finishing a several year run I already have no cclm monthly output files, but only those in directories …/chain/arch and /work/xxx/restarts .
Some of the time series for all months are missing. Now I would like to repeat the post-processing in order to create additional time series of some missing values for each month of my period. Can this be done?
Many thanks for your help
Simon

View in channel
<p> Ok, I understand that you do not have any time series files in work/xxx/post directory. So you can create them from the archived <strong> .tar files. Do you have those? For that you need to untar the archived *.tar files and create time series from the lffd </strong> files stored in the out** directories. For that you can use some of the code in your post(yourEXID).job on them. You probably need the code lines of the post(yourEXID).job below the topics: <ol> <li> set number of boundary lines to be cut off in the time series data <br/> and the code lines below the topic <ol> <li> function for building time series <br/> an the code lines below the topic </li> </ol> </li> <li> —- build time series for selected variables </li> </ol> </p> <p> In the case you have all of your simulated months (the out** directories and lffd* files) in /scratch/…./input/arch and nothing in /scratch/…./input/post, just copy them to /scratch/…./input/post and rerun the postprocessing job for each month. </p> <p> Let me know if that solves your problem now? <br/> Cheers, <br/> Merja </p>

  @merjatölle in #c8b90fc

<p> Ok, I understand that you do not have any time series files in work/xxx/post directory. So you can create them from the archived <strong> .tar files. Do you have those? For that you need to untar the archived *.tar files and create time series from the lffd </strong> files stored in the out** directories. For that you can use some of the code in your post(yourEXID).job on them. You probably need the code lines of the post(yourEXID).job below the topics: <ol> <li> set number of boundary lines to be cut off in the time series data <br/> and the code lines below the topic <ol> <li> function for building time series <br/> an the code lines below the topic </li> </ol> </li> <li> —- build time series for selected variables </li> </ol> </p> <p> In the case you have all of your simulated months (the out** directories and lffd* files) in /scratch/…./input/arch and nothing in /scratch/…./input/post, just copy them to /scratch/…./input/post and rerun the postprocessing job for each month. </p> <p> Let me know if that solves your problem now? <br/> Cheers, <br/> Merja </p>
MT replied

Ok, I understand that you do not have any time series files in work/xxx/post directory. So you can create them from the archived .tar files. Do you have those? For that you need to untar the archived *.tar files and create time series from the lffd files stored in the out** directories. For that you can use some of the code in your post(yourEXID).job on them. You probably need the code lines of the post(yourEXID).job below the topics:

  1. set number of boundary lines to be cut off in the time series data
    and the code lines below the topic
    1. function for building time series
      an the code lines below the topic
  2. —- build time series for selected variables

In the case you have all of your simulated months (the out** directories and lffd* files) in /scratch/…./input/arch and nothing in /scratch/…./input/post, just copy them to /scratch/…./input/post and rerun the postprocessing job for each month.

Let me know if that solves your problem now?
Cheers,
Merja

View in channel
<p> Dear Merja,colleagues </p> <p> I do not have the files unfortunately. The only lffd files I have are those in /chain/arch/*** directory, where some of parameters we need are missing. <br/> The files with results of our simulated months (the out** directories and lffd* files) in /scratch/…./input/arch have been deleted. <br/> So, we will apparently have to re-run our experiment. Please remind me what should be done to keep the intermediate files in out** directories notdeleted. Regards, <br/> Simon </p>

  @redc_migration in #fe22de1

<p> Dear Merja,colleagues </p> <p> I do not have the files unfortunately. The only lffd files I have are those in /chain/arch/*** directory, where some of parameters we need are missing. <br/> The files with results of our simulated months (the out** directories and lffd* files) in /scratch/…./input/arch have been deleted. <br/> So, we will apparently have to re-run our experiment. Please remind me what should be done to keep the intermediate files in out** directories notdeleted. Regards, <br/> Simon </p>
re replied

Dear Merja,colleagues

I do not have the files unfortunately. The only lffd files I have are those in /chain/arch/*** directory, where some of parameters we need are missing.
The files with results of our simulated months (the out** directories and lffd* files) in /scratch/…./input/arch have been deleted.
So, we will apparently have to re-run our experiment. Please remind me what should be done to keep the intermediate files in out** directories notdeleted. Regards,
Simon

View in channel
<p> The files in the archive directory /chain/arch/*** are the same as in /scratch/…./input/arch, i.e. all out*** directories are saved from scratch. They are just put into tar files. If you do not find your parameters in the tar files you have not declared them in the namelist for output (assuming your job runs without failure). <br/> Therefore please make sure that your parameter is <br/> 1. in the namelist output list of <span class="caps"> CCLM </span> (check the cclm.job.tmpl) <br/> 2. in a “timeseries” line in the post.job.tmpl. In case not add a “timeseries” line for your parameter (for examples of such a line look into in post.job.tmpl) </p>

  @burkhardtrockel in #b83ee2a

<p> The files in the archive directory /chain/arch/*** are the same as in /scratch/…./input/arch, i.e. all out*** directories are saved from scratch. They are just put into tar files. If you do not find your parameters in the tar files you have not declared them in the namelist for output (assuming your job runs without failure). <br/> Therefore please make sure that your parameter is <br/> 1. in the namelist output list of <span class="caps"> CCLM </span> (check the cclm.job.tmpl) <br/> 2. in a “timeseries” line in the post.job.tmpl. In case not add a “timeseries” line for your parameter (for examples of such a line look into in post.job.tmpl) </p>
BR replied

The files in the archive directory /chain/arch/*** are the same as in /scratch/…./input/arch, i.e. all out*** directories are saved from scratch. They are just put into tar files. If you do not find your parameters in the tar files you have not declared them in the namelist for output (assuming your job runs without failure).
Therefore please make sure that your parameter is
1. in the namelist output list of CCLM (check the cclm.job.tmpl)
2. in a “timeseries” line in the post.job.tmpl. In case not add a “timeseries” line for your parameter (for examples of such a line look into in post.job.tmpl)

View in channel