This is the bug I'm facing. I deliberately configured a container so that the source of a <filesystem/> to passthrough doesn't exist. The start fails with: lxcContainerPivotRoot:669 : Failed to create /non-existent/path/.oldroot: Permission denied which is expected. But what is NOT expected is that CGroup hierarchy is left behind. This is because the controller sets up the CGroup hierarchy, user namespace, moves interfaces, etc. and finally checks whether container setup (done in a separate process) succeeded. Only after all this the error is propagated to the LXC driver. The driver aborts the startup and tries to perform the cleanup, but this is missing CGroups because those weren't detected yet. Ideally, whenever a function fails, it tries to unroll back so that is has no artifacts left behind (look at all those frees/FD closes/etc. at end of functions). But with CGroups it is different - the controller process can't clean up after itself, because it is still running inside that CGroup. Therefore, what we have to do is to let the driver detect CGroups as soon as they are created, and proceed with controller execution only after that. Signed-off-by: Michal Privoznik <mprivozn@xxxxxxxxxx> --- src/lxc/lxc_controller.c | 19 +++++++++++++++++-- src/lxc/lxc_process.c | 20 ++++++++++++++++++++ 2 files changed, 37 insertions(+), 2 deletions(-) diff --git a/src/lxc/lxc_controller.c b/src/lxc/lxc_controller.c index 1c0a370d4b..33f397c11d 100644 --- a/src/lxc/lxc_controller.c +++ b/src/lxc/lxc_controller.c @@ -348,7 +348,7 @@ static int virLXCControllerConsoleSetNonblocking(virLXCControllerConsole *consol } -static int virLXCControllerDaemonHandshake(virLXCController *ctrl) +static int virLXCControllerDaemonHandshakeCont(virLXCController *ctrl) { if (lxcContainerSendContinue(ctrl->handshakeFds[1]) < 0) { virReportSystemError(errno, "%s", @@ -358,6 +358,15 @@ static int virLXCControllerDaemonHandshake(virLXCController *ctrl) return 0; } +static int virLXCControllerDaemonHandshakeWait(virLXCController *ctrl) +{ + if (lxcContainerWaitForContinue(ctrl->handshakeFds[0]) < 0) { + virReportSystemError(errno, "%s", + _("error waiting for continue signal from daemon")); + return -1; + } + return 0; +} static int virLXCControllerValidateNICs(virLXCController *ctrl) { @@ -2372,6 +2381,11 @@ virLXCControllerRun(virLXCController *ctrl) if (virLXCControllerSetupCgroupLimits(ctrl) < 0) goto cleanup; + /* Allow daemon to detect CGroups. */ + if (virLXCControllerDaemonHandshakeCont(ctrl) < 0 || + virLXCControllerDaemonHandshakeWait(ctrl) < 0) + goto cleanup; + if (virLXCControllerSetupUserns(ctrl) < 0) goto cleanup; @@ -2401,7 +2415,8 @@ virLXCControllerRun(virLXCController *ctrl) if (virLXCControllerConsoleSetNonblocking(&(ctrl->consoles[i])) < 0) goto cleanup; - if (virLXCControllerDaemonHandshake(ctrl) < 0) + /* Allow daemon to connect to the monitor. */ + if (virLXCControllerDaemonHandshakeCont(ctrl) < 0) goto cleanup; /* and preemptively close handshakeFds */ diff --git a/src/lxc/lxc_process.c b/src/lxc/lxc_process.c index 493e19f03d..55e74e913b 100644 --- a/src/lxc/lxc_process.c +++ b/src/lxc/lxc_process.c @@ -1473,6 +1473,7 @@ int virLXCProcessStart(virConnectPtr conn, if (g_atomic_int_add(&driver->nactive, 1) == 0 && driver->inhibitCallback) driver->inhibitCallback(true, driver->inhibitOpaque); + /* The first synchronization point is when the controller creates CGroups. */ if (lxcContainerWaitForContinue(handshakefds[0]) < 0) { char out[1024]; @@ -1504,6 +1505,25 @@ int virLXCProcessStart(virConnectPtr conn, goto cleanup; } + if (lxcContainerSendContinue(handshakefds[3]) < 0) { + virReportSystemError(errno, "%s", + _("Failed to send continue signal to controller")); + goto cleanup; + } + + /* The second synchronization point is when the controller finished + * creating the container. */ + if (lxcContainerWaitForContinue(handshakefds[0]) < 0) { + char out[1024]; + + if (!(virLXCProcessReadLogOutput(vm, logfile, pos, out, 1024) < 0)) { + virReportError(VIR_ERR_INTERNAL_ERROR, + _("guest failed to start: %s"), out); + } + + goto cleanup; + } + /* And we can get the first monitor connection now too */ if (!(priv->monitor = virLXCProcessConnectMonitor(driver, vm))) { /* Intentionally overwrite the real monitor error message, -- 2.26.3