github/kata-containers
1
0
Fork 1
mirror of https://github.com/kata-containers/kata-containers.git synced 2025-04-28 19:54:35 +00:00
Code Issues Projects Releases Wiki Activity

agent: best-effort removing mount point

Browse Source 
During container exit, the agent tries to remove all the mount point directories,
which can fail if it's a readonly filesytem (e.g. device mapper). This commit ignores
the removal failure and logs a warning message.

Fixes: #4043

Signed-off-by: Feng Wang <feng.wang@databricks.com>
This commit is contained in:
Feng Wang 2022-04-11 18:44:23 -07:00
parent c108bb7a2a
commit aabcebbf58
1 changed files with 7 additions and 13 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

20
src/agent/src/sandbox.rs
View File

@ -151,7 +151,12 @@ impl Sandbox {
pub fn remove_sandbox_storage(&self, path: &str) -> Result<()> {
let mounts = vec![path.to_string()];
remove_mounts(&mounts)?;
fs::remove_dir_all(path).context(format!("failed to remove dir {:?}", path))?;
// "remove_dir" will fail if the mount point is backed by a read-only filesystem.
// This is the case with the device mapper snapshotter, where we mount the block device directly
// at the underlying sandbox path which was provided from the base RO kataShared path from the host.
if let Err(err) = fs::remove_dir(path) {
warn!(self.logger, "failed to remove dir {}, {:?}", path, err);
}
Ok(())
}
@ -564,19 +569,8 @@ mod tests {
.remove_sandbox_storage(invalid_dir.to_str().unwrap())
.is_err());
// Now, create a double mount as this guarantees the directory cannot
// be deleted after the first umount.
for _i in 0..2 {
assert!(bind_mount(srcdir_path, destdir_path, &logger).is_ok());
}
assert!(bind_mount(srcdir_path, destdir_path, &logger).is_ok());
assert!(
s.remove_sandbox_storage(destdir_path).is_err(),
"Expect fail as deletion cannot happen due to the second mount."
);
// This time it should work as the previous two calls have undone the double
// mount.
assert!(s.remove_sandbox_storage(destdir_path).is_ok());
}