Scheduled Downtime
On Friday 21 April 2023 @ 5pm MT, this website will be down for maintenance and expected to return online the morning of 24 April 2023 at the latest

A strange wrf.exe behavior

This post was from a previous version of the WRF&MPAS-A Support Forum. New replies have been disabled and if you have follow up questions related to this post, then please start a new thread from the forum home page.

sedbar

New member
Is there anything in REGISTRY files or namelist which could control maximum wfrout size?

I could copy wrfout with 27GB in size from another server to mine, but my WRF model IO on my server hangs when wrfout reach 20GB in size, of course wrf.exe instances still running showed by top command.

I check server setting as I could with a normal account I have, but no quota sets on the server

I am curious what could the reason of such problem. If anyone has any idea I am pleased to share it with me.

P.S. I reduced frames_per_outfile to 12 to have multiple smaller wrfout files.

Thanks
Majid
 
Hello again,
I take a deep look at namelist.input and found that may be "restart_interval" is the main culprit, as you see restart set to .false. and restart_interval was set to 6000 ( I change it to 10000 minutes), at 6000 minutes after wrf starts, its try to write down a restart file to hard disk but: 1) it won't as restart set false , 2) why its IO hangs on when try to write down restart file.

So I change restart_interval to 10000 which more than my wrf runs ( 6 days), but the main question remains: why wrf IO hangs on saving restart file.

Here is part of my namelist.input:

Thanks
Majid


input_from_file = .true.,.true.,.true.,
history_interval =180,60,60,
frames_per_auxhist2 = 1000,1000,
frames_per_outfile = 1000, 1000, 1000,
restart = .false.,
restart_interval = 10000,
io_form_history = 2
io_form_restart = 2
io_form_input = 2
io_form_boundary = 2
debug_level = 0
 
Hi Majid,
Restart files are larger than general wrfout* files. They contain more information required for initiating a restart. There typically is a limit on the size of the files, but this is related to netCDF (netCDF places a limit on file size), and not the WRF model, itself. Setting the restart interval to a longer period is one way to overcome the wrfrst* file write-out, but also setting that to a much shorter period will allow the files to be smaller and will perhaps write out. 20 GB is still very large for a wrfout* file. It may be necessary to cut those down, as well, by setting the frames_per_outfile to a smaller value. You could also use the split-netcdf option (io_form_history - 102), if you're seeing having issues. If you still need additional help, please attach your full namelist.input so I can take a look. Thanks!
 
Top