This patch is not only an optimization. It also allows to stop restored container with NFS but unreachable network, because tasks can be blocked in kernel (due to SPFS "Stub" mode) and can't be frozen. Killing tasks before freezing will kill spfs process, thus either killing all the container tasks without freezing (and optimization), or at least release processes, waiting for spfs reply in kernel.
https://jira.sw.ru/browse/PSBM-68232 Signed-off-by: Stanislav Kinsburskiy <skinsbur...@virtuozzo.com> --- lib/env_nsops.c | 4 ++++ 1 file changed, 4 insertions(+) diff --git a/lib/env_nsops.c b/lib/env_nsops.c index 221fad7..4ecd46e 100644 --- a/lib/env_nsops.c +++ b/lib/env_nsops.c @@ -1110,6 +1110,10 @@ static int ns_env_stop_force(struct vzctl_env_handle *h) logger(0, 0, "Forcibly stop the Container..."); + ret = ns_env_kill(h, 0); + if (ret) + goto release_sunrpc; + ret = cg_freezer_cmd(EID(h), VZCTL_CMD_FREEZE); if (ret) goto release_sunrpc; _______________________________________________ Devel mailing list Devel@openvz.org https://lists.openvz.org/mailman/listinfo/devel
