-
Notifications
You must be signed in to change notification settings - Fork 57
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
Ensure that config files are placed atomically #103
Comments
Funnily enough this was the approach that salt followed and that brought us some headaches when we were writing manifests in While it seems the safest bet, using the target directory to write the tempfile can have unexpected side effects. |
I'm going to call that an error in kublet, given that using a tempname in the target directory is pretty standard practice in all config management tools I'm aware of. :D In that kind of rare situation, doing a stat and choosing a target which has the same device number is really what's important. Any POSIX filesystem guarantees that a |
Does this still apply? I am looking to cleanup all the cruft in issues :) |
Following up. One of us will either close or do some research. |
This is still technically "a thing", but somewhat insignificant given future direction. Closing. |
Per the discussion linked below, "someone" should verify that the file copy process is performing atomic copies. Namely, that it's creating a $tempname on the destination (in the same directory as $realname, so the filesystem is the same and the actual rename syscall can be used), populating the contents to that temp name, and then renaming $tempname to $realname.
Originally posted by @dannysauer in https://github.com/SUSE/caaspctl/pull/98/review_comment/create
(renamed to #98)
The text was updated successfully, but these errors were encountered: