Closed barneydobson closed 4 weeks ago
If it happened on the cluster only, it can be due to using the system tmp
dir when using python's tempfile
. It can fill up quickly on a cluster, so it's best to create a local tmp
directory in your own scratch storage instead of using the system's. You can use the dir
arg in, for example, tempfile.TemporaryDirectory
.
Occasionally the
flow_dir
calculation fails duringclip_to_catchments
. This has only happened to me on a PBS jobarray - so might be something to do with that, unsure.Probably aren't helpful - but logs attached for reference. logs.zip
I think it's a rust thing... which i don't really know about