Can the m.Freeze and m.Thawed be canceled?

[lijinyong]

The lack of error handling for m.Freeze and m.Thawed in the code may cause errors when freezing and thawing cgroups, which may affect the correctness of the program.

For example, when executing m.Freeze, if runc exits abnormally (due to fault injection or other abnormal scenarios), all business processes within the cgroup will be frozen.

// signalAllProcesses freezes then iterates over all the processes inside the
// manager's cgroups sending the signal s to them.
// If s is SIGKILL then it will wait for each process to exit.
// For all other signals it will check if the process is ready to report its
// exit status and only if it is will a wait be performed.
func signalAllProcesses(m cgroups.Manager, s os.Signal) error {
var procs []*os.Process
if err := m.Freeze(configs.Frozen); err != nil {
logrus.Warn(err)
}
pids, err := m.GetAllPids()
if err != nil {
if err := m.Freeze(configs.Thawed); err != nil {
logrus.Warn(err)
}
return err
}
for _, pid := range pids {
p, err := os.FindProcess(pid)
if err != nil {
logrus.Warn(err)
continue
}
procs = append(procs, p)
if err := p.Signal(s); err != nil {
logrus.Warn(err)
}
}
if err := m.Freeze(configs.Thawed); err != nil {
logrus.Warn(err)
}

[lifubang]

I think they can't be canceled, one obvious reason is that it may cause PID reuse attack.
As you say, If runc exits abnormally, we can call runc again to retry. Because signalAllProcesses is used in a separate RUNC command process. The command caller should have detect and retry mechanism.

[lijinyong]

After the container is frozen, runc is not allowed to be executed in a paused container. and all retry repair mechanisms that ultimately rely on runc will of course fail.

Although runc provides a new feature ignore-paused to support the execution of runc in the pause container, containedred seems to have no intention of integration.

[lifubang]

But runc kill is allowed to be executed for a paused container.