forecast error in HWRF V4.0a

Submitted by zahra.ranji on Wed, 03/31/2021 - 11:15
Forum: Users | Forecast

Hi there,

I am trying to run IRMA example on HWRF V4.0. I noticed that there are known issues on running relocation and gsi in this version. So I skipped them by the recommendation provided in https://dtcenter.org/community-code/hurricane-wrf-hwrf/hwrf-system-components-v4-0a#issues

However, this example is stopped by an error in forecast wrapper (the rest of components were successfully completed):

[proxy:0:0@localhost.localdomain] HYDU_create_process (../../utils/launch/launch.c:390): pipe error (Too many open files)
[proxy:0:0@localhost.localdomain] launch_procs (../../pm/pmiserv/pmip_cb.c:2988): create process returned error
[proxy:0:0@localhost.localdomain] HYD_pmcd_pmip_control_cmd_cb (../../pm/pmiserv/pmip_cb.c:3638): launch_procs returned error
[proxy:0:0@localhost.localdomain] HYDT_dmxu_poll_wait_for_event (../../tools/demux/demux_poll.c:76): callback returned error status
[proxy:0:0@localhost.localdomain] main (../../pm/pmiserv/pmip.c:558): demux engine error waiting for event
[mpiexec@localhost.localdomain] control_cb (../../pm/pmiserv/pmiserv_cb.c:834): connection to proxy 0 at host localhost.localdomain failed
[mpiexec@localhost.localdomain] HYDT_dmxu_poll_wait_for_event (../../tools/demux/demux_poll.c:76): callback returned error status
[mpiexec@localhost.localdomain] HYD_pmci_wait_for_completion (../../pm/pmiserv/pmiserv_pmci.c:507): error waiting for event
[mpiexec@localhost.localdomain] main (../../ui/mpich/mpiexec.c:1148): process manager error waiting for completion
03/31 20:20:08.068 hwrf.runwrf (run.py:398) INFO:   - exit status 255
03/31 20:20:08.069 hwrf.runwrf (fcsttask.py:934) INFO: wrf: exit status 255
03/31 20:20:08.069 hwrf.runwrf (fcsttask.py:984) CRITICAL: runwrf failed: [Errno 2] No such file or directory: 'rsl.out.0000'
Traceback (most recent call last):
  File "/home2/xcyclone/source/HWRF/hwrfrun/ush/hwrf/fcsttask.py", line 979, in run
    self.run_exe()
  File "/home2/xcyclone/source/HWRF/hwrfrun/ush/hwrf/coupling.py", line 226, in run_exe
    super(CoupledWRF,self).run_exe('wrf',None,cmd)
  File "/home2/xcyclone/source/HWRF/hwrfrun/ush/hwrf/fcsttask.py", line 1704, in run_exe
    runner=runner,sleeptime=sleeptime)
  File "/home2/xcyclone/source/HWRF/hwrfrun/ush/hwrf/fcsttask.py", line 936, in run_exe
    logger=logger):
  File "/home2/xcyclone/source/HWRF/hwrfrun/ush/produtil/fileop.py", line 908, in check_last_lines
    with open(str(filename),'rt') as f:
IOError: [Errno 2] No such file or directory: 'rsl.out.0000'
03/31 20:20:08Z NONAME-hwrf.runwrf: CRITICAL:  runwrf failed: [Errno 2] No such file or directory: 'rsl.out.0000'
03/31 20:20:08.071 hwrf.runwrf (cd.py:186) WARNING: /home2/xcyclone/HWRF/hwrfrun/2017090512/11L/runwrf: leaving temp directory due to exception
03/31 20:20:08.071 hwrf.runwrf (cd.py:188) INFO: /home2/xcyclone/HWRF/hwrfrun/2017090512/11L/runwrf: listing current directory (/home2/xcyclone/HWRF/hwrfrun/2017090512/11L/runwrf) via ls -l
T='/home2/xcyclone/HWRF/hwrfrun/com/2017090512/11L'

Is this error related to the skipped components? Should I downgrade HWRF to tun this example?

Thanks in advance for your help.

Permalink

Submitted by biswas on Wed, 03/31/2021 - 11:41

Hi,

The issue listed pertains to running HWRFv4.0a with the most recent FV3 based GFS upgrades. The input file formats have changed (from nemsio to NetCDF and upgrades to bufrlib files). If you are running with the dataset provided with the release you can run the full configuration. However, the reduced capability should also run with the Irma case. It is up to you which configuration is better for your study. The error messages don't say much except that it can't find the rsl file which means the model didn't start at all. Please check whether all the previous steps ran fine. You need to run only these wrappers for the reduced capability:

1. launcher_wrapper

2. init_gfs_wrapper

init_ocean_wrapper

3. init_bdy_wrapper

4. unpost_wrapper

5. forecast_wrapper

post_wrapper

products_wrapper

How many processors are you using to run the forecast_wrapper? Please refer to the HWRF Users Guide for recommended procs count to run the task. Resubmit the task and see if you get the same error. If this doesn't fix the problem I may need to look at all the log files and storm1.conf. 

Thanks

Biswas

Permalink

Submitted by zahra.ranji on Thu, 04/01/2021 - 03:04

Thank you Biswas for your quick response,

I am trying to run this example to know more about the model structure. Then, I want to move to the Arabian Sea and at that time, I can get access to an HPC with required configuration.

Since I am running this example on a PC with 8 processors and 32 GB memory, the recommended processors in HWRF manual (4 for the coupler and 9 for POM) are not provided for me.  

Above sets of wrappers, together with relocate, suit my study area (i.e., North Indian Ocean). I had to change the domain and decrease the resolution for 5 times (!) and then all the wrappers works perfectly and ended with "successful completion" and all the outputs are created in the related directory. But I have still this problem in forecast component.

In this case, how can I face with this problem? 

I am so grateful for your helps and time.

Permalink

Submitted by zahra.ranji on Tue, 04/13/2021 - 00:30

Hi,

For the users who face with similar problem, I should say this issue is solved by modifying nproc_x and nproc_y in storm1.conf file, as mentioned in HWRF Manual.

Regards,

Zahra

Sitemap

Contact

303.497.8921
dtcenter@ucar.edu

Addresses

mailing:
PO Box 3000
Boulder, CO 80307-3000

Shipping:
3090 Center Green Dr.
Boulder, CO 80301

Follow Us

YouTube Icon
Twitter Icon

Copyright © 2024. All rights reserved.