LIS Testcases DA vague error #896
-
|
Hello, our lab has been working to incorporate LIS into our modeling. We have successfully run standard LIS at a variety of resolutions. I have been attempting to work through the data assimilation test cases in order to understand the process better so that we can incorporate it into our work. Okay, now for the error. Running on NASA Discover, I have successfully worked through test cases 1 through 5. While working through testcase 6, I encounter a vague error. Consulting the log files, it appears that every piece of input gets successfully read by LIS. The first error is a warning that states [WARN] DA obs domain details The second error is denoted with [ERR] [ERR] SMAP(NASA) soil moisture Composite Release ID: is missing Stopping. Has anyone experienced these errors before? If so, has anyone found a solution to these errors? Best Jack |
Beta Was this translation helpful? Give feedback.
Replies: 6 comments 6 replies
-
|
Hello Jack, The first warning should actually appear in your log files as: [INFO] DA obs domain details: If you are not seeing exactly what I have above, please let us know. This warning is not a concern - it's simply letting the user know that the DA obs domain netCDF file does not have the GRIDTYPE attribute. This warning will not affect the run (although the LIS team should fix this to ensure that the DA obs domain files are written with this attribute). The second error is stopping the run. The testcase suite is designed to work with the specific GitHub commit ID on the bottom of this page: You may be using a more recent or current version of the software, which is looking for this missing config entry in the lis.config file. New code added since that commit ID is looking for this config entry. If you look at the LIS Users' Guide (best viewed in the browser as detail on this page): https://github.com/NASA-LIS/LISF/tree/master/docs You can see that there is an explanation of this config entry. Please add it to your lis.config and try again. You may be using R17 instead of R16 - please check the filenames of the data you are using on Discover. SMAP(NASA) soil moisture Composite Release ID: specifies the SMAP NASA composite release ID included within the filename (e.g., R16). David |
Beta Was this translation helpful? Give feedback.
-
|
I am able to run step 6 of the public testcase with LIS compiled from the current master branch. We updated the testcase files last summer which involved the addition of the @dechowjack, can you check that you are using the latest version of the testcase files? The latest version of the tar files end in As @dmocko mentioned, you'll have the best luck running the public testcase with LDT/LIS/LVT compiled from the commit indicated on the testcase website. See the "Setup" section in this testcase walkthrough doc I created for instructions: https://github.com/bmcandr/lis-public-tc-walkthrough |
Beta Was this translation helpful? Give feedback.
-
|
@dmocko Hi David, the first warning does indeed appear that way in my log files, I just left out the warnings when I typed it up. Second, I went into the config file and the composite ID is specified as R16. When I look at the SMAP files, they have the name format of SMAP_L3_SM_P_20170301_R16010_001.h5 Does this mean that the ID is actually 16.xxx? Or should the R16 ID be enough? @bmcandr Hi Brandon, I am indeed working from the 2020 gzip files. |
Beta Was this translation helpful? Give feedback.
-
|
@dechowjack You should just use "R16" (without the quotes) for this config entry. You may have a typo, or extra space, or change in capitalization in your lis.config file. Can you please double check? LIS is unable to find this exact phrase in the lis.config file:SMAP(NASA) soil moisture Composite Release ID: R16David |
Beta Was this translation helpful? Give feedback.
-
|
Sorry for the weird random bolding. I hate copy and paste most of the time.... |
Beta Was this translation helpful? Give feedback.
-
|
I just checked, it is identical to the line you posted. |
Beta Was this translation helpful? Give feedback.
-
Beta Was this translation helpful? Give feedback.
-
|
@dmocko I've been working on installing/compiling the LIS framework with the same commit ID as the public testcases so I can rerun the test cases. While doing so, I ran into an error installing one of the dependencies (OpenJPG) requiring sudo access to finish the install. I have tried to install in both home and $nobackup directory. Do you have any suggestions on what to do? |
Beta Was this translation helpful? Give feedback.
-
|
@dechowjack, you mentioned that you're working on Discover. We have all of the necessary libraries already installed and you should be able to use the module files under I have some instructions here for setting up your environment on Discover: https://github.com/bmcandr/lis-bootcamp/blob/master/docs/LIS-on-NCCS-discover.md#setting-up-your-environment-on-discover |
Beta Was this translation helpful? Give feedback.
-
|
Thanks, @bmcandr - I was just going to suggest the same. @dechowjack - I e-mailed you yesterday at your OSU address about you sharing your Discover directory, so I can take a look to see why you are getting that error. Did you get that e-mail? David |
Beta Was this translation helpful? Give feedback.
-
|
@dechowjack - Thanks for sharing your directory to me via e-mail. I can't see anything wrong with it, so it may be that your libraries are causing a conflict. Can you please try the modulefile setup method, and use only those libraries to compile? Be sure you don't have other modules loaded when you compile/run, as this can cause a conflict. |
Beta Was this translation helpful? Give feedback.
-
|
Hi All, After recompiling the most recent commit of LIS, I was able to successfully run the data assimilation test case with all of my previous inputs. The only thing that I changed was the compiled LIS executable in the folder. Thanks for all the help. Best Jack |
Beta Was this translation helpful? Give feedback.
@dmocko @bmcandr @LIS-navari
Hi All,
After recompiling the most recent commit of LIS, I was able to successfully run the data assimilation test case with all of my previous inputs. The only thing that I changed was the compiled LIS executable in the folder. Thanks for all the help.
Best
Jack