github/kata-containers
1
0
Fork 2
mirror of https://github.com/kata-containers/kata-containers.git synced 2026-07-02 07:02:16 +00:00
Code Issues Projects Releases Wiki Activity

runtime-rs: make sandbox cgroup runtime attach idempotent

Browse Source 
The dragonball nerdctl CI job can race when creating and attaching the
runtime process to the sandbox cgroup, surfacing an os error 17
(AlreadyExists) during shim task creation.

Let's retry add_proc once on this pre-existing cgroup condition so
startup remains robust.

Signed-off-by: Fabiano Fidêncio <ffidencio@nvidia.com>
Assisted-by: Codex <codex@openai.com>
This commit is contained in:
Fabiano Fidêncio
2026-06-07 13:38:13 +02:00
parent 4d569c22b4
commit 4dc288401e
1 changed files with 41 additions and 6 deletions
Download Patch File Download Diff File Expand all files Collapse all files

47
src/runtime-rs/crates/resource/src/cgroups/resource_inner.rs
View File

@@ -5,6 +5,7 @@
//
use std::collections::{HashMap, HashSet};
use std::error::Error as _;
use std::process;
use std::str::FromStr;
use std::time::Duration;
@@ -41,6 +42,36 @@ pub(crate) struct CgroupsResourceInner {
}
impl CgroupsResourceInner {
fn is_already_exists_error(err: &cgroups_rs::manager::Error) -> bool {
let mut source = err.source();
while let Some(inner_err) = source {
if let Some(io_err) = inner_err.downcast_ref::<std::io::Error>() {
if io_err.kind() == std::io::ErrorKind::AlreadyExists {
return true;
}
}
source = inner_err.source();
}
false
}
fn add_proc_with_existing_retry(
cgroup: &mut CgroupManager,
pid: CgroupPid,
context: &str,
) -> Result<()> {
match cgroup.add_proc(pid) {
Ok(_) => Ok(()),
Err(err) if Self::is_already_exists_error(&err) => cgroup
.add_proc(pid)
.with_context(|| format!("{context} (retry after pre-existing cgroup)")),
Err(err) => Err(err).context(context.to_string()),
}
}
/// Create cgroup managers according to the cgroup configuration.
///
/// # Returns
@@ -90,13 +121,17 @@ impl CgroupsResourceInner {
// The runtime is prioritized to be added to the overhead cgroup.
let pid = CgroupPid::from(process::id() as u64);
if let Some(overhead_cgroup) = overhead_cgroup.as_mut() {
overhead_cgroup
.add_proc(pid)
.context("add runtime to overhead cgroup")?;
Self::add_proc_with_existing_retry(
overhead_cgroup,
pid,
"add runtime to overhead cgroup",
)?;
} else {
sandbox_cgroup
.add_proc(pid)
.context("add runtime to sandbox cgroup")?;
Self::add_proc_with_existing_retry(
&mut sandbox_cgroup,
pid,
"add runtime to sandbox cgroup",
)?;
}
Ok(Self {