[libvirt] [PATCH] Keep track of guest paused state after disk IO / watchdog events

[Daniel P. Berrange]

When a watchdog/IO error occurs, one of the possible actions that
QEMU might take is to pause the guest. In this scenario libvirt
needs to update its internal state for the VM, and emit a
lifecycle event:

  VIR_DOMAIN_EVENT_SUSPENDED

with a detail being one of:

  VIR_DOMAIN_EVENT_SUSPENDED_IOERROR
  VIR_DOMAIN_EVENT_SUSPENDED_WATCHDOG

To future proof against possible QEMU support for multiple monitor
consoles, this patch also hooks into the 'STOPPED' event in QEMU
and emits a generic VIR_DOMAIN_EVENT_SUSPENDED_PAUSED event

* include/libvirt/libvirt.h.in: Add VIR_DOMAIN_EVENT_SUSPENDED_IOERROR
* src/qemu/qemu_driver.c: Update VM state to paused when IO error
  or watchdog events occurrs
* src/qemu/qemu_monitor_json.c: Fix typo in disk IO event name
---
 include/libvirt/libvirt.h.in |    2 +
 src/qemu/qemu_driver.c       |   83 ++++++++++++++++++++++++++++++++++++++----
 src/qemu/qemu_monitor_json.c |    2 +-
 3 files changed, 78 insertions(+), 9 deletions(-)

diff --git a/include/libvirt/libvirt.h.in b/include/libvirt/libvirt.h.in
index 6d8552f..7cb483e 100644
--- a/include/libvirt/libvirt.h.in
+++ b/include/libvirt/libvirt.h.in
@@ -1384,6 +1384,8 @@ typedef enum {
 typedef enum {
     VIR_DOMAIN_EVENT_SUSPENDED_PAUSED = 0,   /* Normal suspend due to admin pause */
     VIR_DOMAIN_EVENT_SUSPENDED_MIGRATED = 1, /* Suspended for offline migration */
+    VIR_DOMAIN_EVENT_SUSPENDED_IOERROR = 2,  /* Suspended due to a disk I/O error */
+    VIR_DOMAIN_EVENT_SUSPENDED_WATCHDOG = 3,  /* Suspended due to a watchdog firing */
 } virDomainEventSuspendedDetailType;
 
 /**
diff --git a/src/qemu/qemu_driver.c b/src/qemu/qemu_driver.c
index 9ee5da7..4291bc7 100644
--- a/src/qemu/qemu_driver.c
+++ b/src/qemu/qemu_driver.c
@@ -909,6 +909,38 @@ qemuHandleDomainReset(qemuMonitorPtr mon ATTRIBUTE_UNUSED,
 
 
 static int
+qemuHandleDomainStop(qemuMonitorPtr mon ATTRIBUTE_UNUSED,
+                     virDomainObjPtr vm)
+{
+    struct qemud_driver *driver = qemu_driver;
+    virDomainEventPtr event = NULL;
+
+    virDomainObjLock(vm);
+    if (vm->state == VIR_DOMAIN_RUNNING) {
+        VIR_DEBUG("Transitioned guest %s to paused state due to unknown event", vm->def->name);
+
+        vm->state = VIR_DOMAIN_PAUSED;
+        event = virDomainEventNewFromObj(vm,
+                                         VIR_DOMAIN_EVENT_SUSPENDED,
+                                         VIR_DOMAIN_EVENT_SUSPENDED_PAUSED);
+
+        if (virDomainSaveStatus(driver->caps, driver->stateDir, vm) < 0)
+            VIR_WARN("Unable to save status on vm %s after IO error", vm->def->name);
+    }
+    virDomainObjUnlock(vm);
+
+    if (event) {
+        qemuDriverLock(driver);
+        if (event)
+            qemuDomainEventQueue(driver, event);
+        qemuDriverUnlock(driver);
+    }
+
+    return 0;
+}
+
+
+static int
 qemuHandleDomainRTCChange(qemuMonitorPtr mon ATTRIBUTE_UNUSED,
                           virDomainObjPtr vm,
                           long long offset)
@@ -943,15 +975,32 @@ qemuHandleDomainWatchdog(qemuMonitorPtr mon ATTRIBUTE_UNUSED,
                          int action)
 {
     struct qemud_driver *driver = qemu_driver;
-    virDomainEventPtr event;
+    virDomainEventPtr watchdogEvent = NULL;
+    virDomainEventPtr lifecycleEvent = NULL;
 
     virDomainObjLock(vm);
-    event = virDomainEventWatchdogNewFromObj(vm, action);
+    watchdogEvent = virDomainEventWatchdogNewFromObj(vm, action);
+
+    if (action == VIR_DOMAIN_EVENT_WATCHDOG_PAUSE &&
+        vm->state == VIR_DOMAIN_RUNNING) {
+        VIR_DEBUG("Transitioned guest %s to paused state due to watchdog", vm->def->name);
+
+        vm->state = VIR_DOMAIN_PAUSED;
+        lifecycleEvent = virDomainEventNewFromObj(vm,
+                                                  VIR_DOMAIN_EVENT_SUSPENDED,
+                                                  VIR_DOMAIN_EVENT_SUSPENDED_WATCHDOG);
+
+        if (virDomainSaveStatus(driver->caps, driver->stateDir, vm) < 0)
+            VIR_WARN("Unable to save status on vm %s after IO error", vm->def->name);
+    }
     virDomainObjUnlock(vm);
 
-    if (event) {
+    if (watchdogEvent || lifecycleEvent) {
         qemuDriverLock(driver);
-        qemuDomainEventQueue(driver, event);
+        if (watchdogEvent)
+            qemuDomainEventQueue(driver, watchdogEvent);
+        if (lifecycleEvent)
+            qemuDomainEventQueue(driver, lifecycleEvent);
         qemuDriverUnlock(driver);
     }
 
@@ -966,7 +1015,8 @@ qemuHandleDomainIOError(qemuMonitorPtr mon ATTRIBUTE_UNUSED,
                         int action)
 {
     struct qemud_driver *driver = qemu_driver;
-    virDomainEventPtr event;
+    virDomainEventPtr ioErrorEvent = NULL;
+    virDomainEventPtr lifecycleEvent = NULL;
     const char *srcPath;
     const char *devAlias;
     virDomainDiskDefPtr disk;
@@ -982,12 +1032,28 @@ qemuHandleDomainIOError(qemuMonitorPtr mon ATTRIBUTE_UNUSED,
         devAlias = "";
     }
 
-    event = virDomainEventIOErrorNewFromObj(vm, srcPath, devAlias, action);
+    ioErrorEvent = virDomainEventIOErrorNewFromObj(vm, srcPath, devAlias, action);
+
+    if (action == VIR_DOMAIN_EVENT_IO_ERROR_PAUSE &&
+        vm->state == VIR_DOMAIN_RUNNING) {
+        VIR_DEBUG("Transitioned guest %s to paused state due to IO error", vm->def->name);
+
+        vm->state = VIR_DOMAIN_PAUSED;
+        lifecycleEvent = virDomainEventNewFromObj(vm,
+                                                  VIR_DOMAIN_EVENT_SUSPENDED,
+                                                  VIR_DOMAIN_EVENT_SUSPENDED_IOERROR);
+
+        if (virDomainSaveStatus(driver->caps, driver->stateDir, vm) < 0)
+            VIR_WARN("Unable to save status on vm %s after IO error", vm->def->name);
+    }
     virDomainObjUnlock(vm);
 
-    if (event) {
+    if (ioErrorEvent || lifecycleEvent) {
         qemuDriverLock(driver);
-        qemuDomainEventQueue(driver, event);
+        if (ioErrorEvent)
+            qemuDomainEventQueue(driver, ioErrorEvent);
+        if (lifecycleEvent)
+            qemuDomainEventQueue(driver, lifecycleEvent);
         qemuDriverUnlock(driver);
     }
 
@@ -1090,6 +1156,7 @@ no_memory:
 static qemuMonitorCallbacks monitorCallbacks = {
     .eofNotify = qemuHandleMonitorEOF,
     .diskSecretLookup = findVolumeQcowPassphrase,
+    .domainStop = qemuHandleDomainStop,
     .domainReset = qemuHandleDomainReset,
     .domainRTCChange = qemuHandleDomainRTCChange,
     .domainWatchdog = qemuHandleDomainWatchdog,
diff --git a/src/qemu/qemu_monitor_json.c b/src/qemu/qemu_monitor_json.c
index 3901d46..eac3aca 100644
--- a/src/qemu/qemu_monitor_json.c
+++ b/src/qemu/qemu_monitor_json.c
@@ -66,7 +66,7 @@ struct {
     { "STOP", qemuMonitorJSONHandleStop, },
     { "RTC_CHANGE", qemuMonitorJSONHandleRTCChange, },
     { "WATCHDOG", qemuMonitorJSONHandleWatchdog, },
-    { "DISK_IO_ERROR", qemuMonitorJSONHandleIOError, },
+    { "BLOCK_IO_ERROR", qemuMonitorJSONHandleIOError, },
     { "VNC_CONNECTED", qemuMonitorJSONHandleVNCConnect, },
     { "VNC_INITIALIZED", qemuMonitorJSONHandleVNCInitialize, },
     { "VNC_DISCONNECTED", qemuMonitorJSONHandleVNCDisconnect, },
-- 
1.6.2.5