-
Notifications
You must be signed in to change notification settings - Fork 1.1k
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
Inconsistent symlink following behavior in checkpoint/restart #21134
Comments
Ok, I think it is time to push for This issue can be solved easily by replacing I'll go ahead and put up a PR to make this small change and see if/where our minimal compiler targets fail... |
Ugh |
We're at gcc7 and clang5, right? |
Ok, in the framework tests I only get two failing tests, and both cases are non matching expected errors, where the error message actually improved.
and
|
Ah yeah.... I see you're talking to yourself here about the problem I discussed in #21137 |
Reason
MOOSE's "paths relative to the input file" logic, follows symlinks. I.e. if I have two directories
a
andb
andthe file in
b
is a symlink back toa
. And I want to run two different simulation in those dirs (let's say with CLI args to modify parameters)...If I'm using restart, MOOSE looks for
a/input_out_cp/*
in both cases. Even though theb
simulation CheckpointIO writes tob/input_out_cp/*...
Design
One part of the code seems to apply
realPath
to follow symlinks, while the other does not. Symlinking inputs to different directories to make it easy to keep multiple outputs seems like a very reasonable thing to do. And at the very least the behavior should be consistent.I propose not following the symlinks.
Impact
Consistent restart behavior w.r.t. symlinks.
The text was updated successfully, but these errors were encountered: