WRF v4 reanalysis runs dying
- Thread starter RCarpenter
- Start date
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.
RCarpenter
Member
I'm afraid I don't have any new information for you. The runs always die without giving any kind of error message. The job keeps running though.
Now I am unable to reliably reproduce a crash. Sometimes a restarted run will die at the same point, sometimes not. Normally this might indicate some sort of hardware issue. But I don't think that's the case because it only happens with a certain physics setting, and not at a history dump time. It's also been reproduced on a different system.
To recap:
I wish I could give you more to go on.
Now I am unable to reliably reproduce a crash. Sometimes a restarted run will die at the same point, sometimes not. Normally this might indicate some sort of hardware issue. But I don't think that's the case because it only happens with a certain physics setting, and not at a history dump time. It's also been reproduced on a different system.
To recap:
- This problem happens only with sf_sfclay_physics=1, and it always happens with that setting. It never happens if it is set to something else.
- The runs always die without giving any error message. The job keep running but does not produce any further output.
- This problem occurs in 3.9.1, 4.0, and 4.0.1.
- These are long-running nested grid runs with 3 grids. The failures occur after 5 to 30 days of integration.
- The crashes happen between history dumps, so it does not seem to be an I/O issue.
I wish I could give you more to go on.