-
Notifications
You must be signed in to change notification settings - Fork 39
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
[BUG] geo2rdr
fails to converge on Track 106
#583
Comments
Likely related to midnight crossing. |
It appears the fix for ensuring success of geo2rdr for azimuth time interpolation worked in ISCE2 but not ISCE3 (and therefore the raider step). All the failures are concentrated track 106 Here is log of output
Below is a sample of 10 input granules:
1,975 jobs failed related to this. |
Just double checked these job failures - all are on track 106. Some that were previously grouped in this were on track 91 but all these succeeded. For reference, here are a few.
|
geo2rdr
fails to converge on track 106.geo2rdr
fails to converge on Track 106
Given the wonkiness of Track 106 - will also want to ensure the convergence on these jobs that never got to RAiDER: ACCESS-Cloud-Based-InSAR/DockerizedTopsApp#152 |
@jhkennedy - I am struggling with this; fairly confident it's an environment/build issue along with isce3. Would love your input on how to proceed because 😵💫. Here is a sample product link (see the last of the most recent batch above): https://hyp3-tibet-jpl-contentbucket-81rn23hp7ppf.s3.us-west-2.amazonaws.com/5b7ad51d-89a3-4fea-8abf-de23a060c39c/S1-GUNW-A-R-106-tops-20230323_20220316-225951-00078W_00041N-PP-3769-v3_0_0.nc The relevant command I am using is:
Here are two different scenarios (maybe 2.5 as you will see) that I tried:
Here are the environments. env-conda-docker.txt Upon first inspection, I see that our container uses |
Furthermore, if I try to install the latest RAiDER on the linux server without docker, I run into dependency heck.
Again, this is reinstalling the environment. Although this is using the You can see in this particular environment that we have isce3 at |
On track 106, where each input granule in ref is being given different orbits, our application of
geo2rdr
in ISCE2 and ISCE3 is having issues (in ISCE3, a failure to converge).Here are some sample inputs and logs:
The link to the relevant input GUNW is here.
Please see ACCESS-Cloud-Based-InSAR/DockerizedTopsApp#148 for more details for this highly related problem.
I suspect that deduplication of orbit files via builtin
set
is acting differently for each step. As we show in previous notebook, one of the orbit files is able to successfully complete an azimuth time grid generation, the other not.We will investigate this further on Monday. It amounts to figuring out how to filter out orbit files for the application of our azimuth time grid generation. We will have to discuss with individuals linked on related issue ticket. Or maybe it requires better initialization of the orbit object in ISCE.
The text was updated successfully, but these errors were encountered: