[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
[PATCH v4 17/21] migration/savevm: don't worry if bitmap migration postc
From: |
Vladimir Sementsov-Ogievskiy |
Subject: |
[PATCH v4 17/21] migration/savevm: don't worry if bitmap migration postcopy failed |
Date: |
Mon, 27 Jul 2020 22:42:32 +0300 |
First, if only bitmaps postcopy is enabled (and not ram postcopy)
postcopy_pause_incoming crashes on an assertion
assert(mis->to_src_file).
And anyway, bitmaps postcopy is not prepared to be somehow recovered.
The original idea instead is that if bitmaps postcopy failed, we just
lose some bitmaps, which is not critical. So, on failure we just need
to remove unfinished bitmaps and guest should continue execution on
destination.
Signed-off-by: Vladimir Sementsov-Ogievskiy <vsementsov@virtuozzo.com>
Reviewed-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Reviewed-by: Andrey Shinkevich <andrey.shinkevich@virtuozzo.com>
Reviewed-by: Eric Blake <eblake@redhat.com>
---
migration/savevm.c | 37 ++++++++++++++++++++++++++++++++-----
1 file changed, 32 insertions(+), 5 deletions(-)
diff --git a/migration/savevm.c b/migration/savevm.c
index 45c9dd9d8a..a843d202b5 100644
--- a/migration/savevm.c
+++ b/migration/savevm.c
@@ -1813,6 +1813,9 @@ static void *postcopy_ram_listen_thread(void *opaque)
MigrationIncomingState *mis = migration_incoming_get_current();
QEMUFile *f = mis->from_src_file;
int load_res;
+ MigrationState *migr = migrate_get_current();
+
+ object_ref(OBJECT(migr));
migrate_set_state(&mis->state, MIGRATION_STATUS_ACTIVE,
MIGRATION_STATUS_POSTCOPY_ACTIVE);
@@ -1839,11 +1842,24 @@ static void *postcopy_ram_listen_thread(void *opaque)
trace_postcopy_ram_listen_thread_exit();
if (load_res < 0) {
- error_report("%s: loadvm failed: %d", __func__, load_res);
qemu_file_set_error(f, load_res);
- migrate_set_state(&mis->state, MIGRATION_STATUS_POSTCOPY_ACTIVE,
- MIGRATION_STATUS_FAILED);
- } else {
+ dirty_bitmap_mig_cancel_incoming();
+ if (postcopy_state_get() == POSTCOPY_INCOMING_RUNNING &&
+ !migrate_postcopy_ram() && migrate_dirty_bitmaps())
+ {
+ error_report("%s: loadvm failed during postcopy: %d. All states "
+ "are migrated except dirty bitmaps. Some dirty "
+ "bitmaps may be lost, and present migrated dirty "
+ "bitmaps are correctly migrated and valid.",
+ __func__, load_res);
+ load_res = 0; /* prevent further exit() */
+ } else {
+ error_report("%s: loadvm failed: %d", __func__, load_res);
+ migrate_set_state(&mis->state, MIGRATION_STATUS_POSTCOPY_ACTIVE,
+ MIGRATION_STATUS_FAILED);
+ }
+ }
+ if (load_res >= 0) {
/*
* This looks good, but it's possible that the device loading in the
* main thread hasn't finished yet, and so we might not be in 'RUN'
@@ -1879,6 +1895,8 @@ static void *postcopy_ram_listen_thread(void *opaque)
mis->have_listen_thread = false;
postcopy_state_set(POSTCOPY_INCOMING_END);
+ object_unref(OBJECT(migr));
+
return NULL;
}
@@ -2437,6 +2455,8 @@ static bool
postcopy_pause_incoming(MigrationIncomingState *mis)
{
trace_postcopy_pause_incoming();
+ assert(migrate_postcopy_ram());
+
/* Clear the triggered bit to allow one recovery */
mis->postcopy_recover_triggered = false;
@@ -2521,15 +2541,22 @@ out:
if (ret < 0) {
qemu_file_set_error(f, ret);
+ /* Cancel bitmaps incoming regardless of recovery */
+ dirty_bitmap_mig_cancel_incoming();
+
/*
* If we are during an active postcopy, then we pause instead
* of bail out to at least keep the VM's dirty data. Note
* that POSTCOPY_INCOMING_LISTENING stage is still not enough,
* during which we're still receiving device states and we
* still haven't yet started the VM on destination.
+ *
+ * Only RAM postcopy supports recovery. Still, if RAM postcopy is
+ * enabled, canceled bitmaps postcopy will not affect RAM postcopy
+ * recovering.
*/
if (postcopy_state_get() == POSTCOPY_INCOMING_RUNNING &&
- postcopy_pause_incoming(mis)) {
+ migrate_postcopy_ram() && postcopy_pause_incoming(mis)) {
/* Reset f to point to the newly created channel */
f = mis->from_src_file;
goto retry;
--
2.21.0
- [PATCH v4 08/21] migration/block-dirty-bitmap: rename state structure types, (continued)
- [PATCH v4 08/21] migration/block-dirty-bitmap: rename state structure types, Vladimir Sementsov-Ogievskiy, 2020/07/27
- [PATCH v4 09/21] migration/block-dirty-bitmap: rename dirty_bitmap_mig_cleanup, Vladimir Sementsov-Ogievskiy, 2020/07/27
- [PATCH v4 10/21] migration/block-dirty-bitmap: move mutex init to dirty_bitmap_mig_init, Vladimir Sementsov-Ogievskiy, 2020/07/27
- [PATCH v4 12/21] migration/block-dirty-bitmap: rename finish_lock to just lock, Vladimir Sementsov-Ogievskiy, 2020/07/27
- [PATCH v4 14/21] migration/block-dirty-bitmap: keep bitmap state for all bitmaps, Vladimir Sementsov-Ogievskiy, 2020/07/27
- [PATCH v4 15/21] migration/block-dirty-bitmap: relax error handling in incoming part, Vladimir Sementsov-Ogievskiy, 2020/07/27
- [PATCH v4 11/21] migration/block-dirty-bitmap: refactor state global variables, Vladimir Sementsov-Ogievskiy, 2020/07/27
- [PATCH v4 13/21] migration/block-dirty-bitmap: simplify dirty_bitmap_load_complete, Vladimir Sementsov-Ogievskiy, 2020/07/27
- [PATCH v4 16/21] migration/block-dirty-bitmap: cancel migration on shutdown, Vladimir Sementsov-Ogievskiy, 2020/07/27
- [PATCH v4 17/21] migration/savevm: don't worry if bitmap migration postcopy failed,
Vladimir Sementsov-Ogievskiy <=
- [PATCH v4 18/21] qemu-iotests/199: prepare for new test-cases addition, Vladimir Sementsov-Ogievskiy, 2020/07/27
- [PATCH v4 20/21] qemu-iotests/199: add early shutdown case to bitmaps postcopy, Vladimir Sementsov-Ogievskiy, 2020/07/27
- [PATCH v4 19/21] qemu-iotests/199: check persistent bitmaps, Vladimir Sementsov-Ogievskiy, 2020/07/27
- [PATCH v4 21/21] qemu-iotests/199: add source-killed case to bitmaps postcopy, Vladimir Sementsov-Ogievskiy, 2020/07/27
- Re: [PATCH v4 for-5.1 00/21] Fix error handling during bitmap postcopy, Eric Blake, 2020/07/27