From: "Dr. David Alan Gilbert (git)" <dgilbert@redhat.com>
To: qemu-devel@nongnu.org
Cc: quintela@redhat.com, yury-kotov@yandex-team.ru,
richardw.yang@linux.intel.com, marcandre.lureau@redhat.com,
ivanren@tencent.com
Subject: [Qemu-devel] [PULL 24/33] migration: always initialise ram_counters for a new migration
Date: Thu, 15 Aug 2019 17:34:55 +0100 [thread overview]
Message-ID: <20190815163504.18937-25-dgilbert@redhat.com> (raw)
In-Reply-To: <20190815163504.18937-1-dgilbert@redhat.com>
From: Ivan Ren <ivanren@tencent.com>
This patch fix a multifd migration bug in migration speed calculation, this
problem can be reproduced as follows:
1. start a vm and give a heavy memory write stress to prevent the vm be
successfully migrated to destination
2. begin a migration with multifd
3. migrate for a long time [actually, this can be measured by transferred bytes]
4. migrate cancel
5. begin a new migration with multifd, the migration will directly run into
migration_completion phase
Reason as follows:
Migration update bandwidth and s->threshold_size in function
migration_update_counters after BUFFER_DELAY time:
current_bytes = migration_total_bytes(s);
transferred = current_bytes - s->iteration_initial_bytes;
time_spent = current_time - s->iteration_start_time;
bandwidth = (double)transferred / time_spent;
s->threshold_size = bandwidth * s->parameters.downtime_limit;
In multifd migration, migration_total_bytes function return
qemu_ftell(s->to_dst_file) + ram_counters.multifd_bytes.
s->iteration_initial_bytes will be initialized to 0 at every new migration,
but ram_counters is a global variable, and history migration data will be
accumulated. So if the ram_counters.multifd_bytes is big enough, it may lead
pending_size >= s->threshold_size become false in migration_iteration_run
after the first migration_update_counters.
Signed-off-by: Ivan Ren <ivanren@tencent.com>
Reviewed-by: Juan Quintela <quintela@redhat.com>
Reviewed-by: Wei Yang <richardw.yang@linux.intel.com>
Suggested-by: Wei Yang <richardw.yang@linux.intel.com>
Message-Id: <1564741121-1840-1-git-send-email-ivanren@tencent.com>
Signed-off-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
---
migration/migration.c | 25 +++++++++++++++++++------
migration/savevm.c | 1 +
2 files changed, 20 insertions(+), 6 deletions(-)
diff --git a/migration/migration.c b/migration/migration.c
index 12b8e5dbe5..c49e9dc035 100644
--- a/migration/migration.c
+++ b/migration/migration.c
@@ -1911,6 +1911,11 @@ static bool migrate_prepare(MigrationState *s, bool blk, bool blk_inc,
}
migrate_init(s);
+ /*
+ * set ram_counters memory to zero for a
+ * new migration
+ */
+ memset(&ram_counters, 0, sizeof(ram_counters));
return true;
}
@@ -3034,6 +3039,17 @@ static void migration_calculate_complete(MigrationState *s)
}
}
+static void update_iteration_initial_status(MigrationState *s)
+{
+ /*
+ * Update these three fields at the same time to avoid mismatch info lead
+ * wrong speed calculation.
+ */
+ s->iteration_start_time = qemu_clock_get_ms(QEMU_CLOCK_REALTIME);
+ s->iteration_initial_bytes = migration_total_bytes(s);
+ s->iteration_initial_pages = ram_get_total_transferred_pages();
+}
+
static void migration_update_counters(MigrationState *s,
int64_t current_time)
{
@@ -3069,9 +3085,7 @@ static void migration_update_counters(MigrationState *s,
qemu_file_reset_rate_limit(s->to_dst_file);
- s->iteration_start_time = current_time;
- s->iteration_initial_bytes = current_bytes;
- s->iteration_initial_pages = ram_get_total_transferred_pages();
+ update_iteration_initial_status(s);
trace_migrate_transferred(transferred, time_spent,
bandwidth, s->threshold_size);
@@ -3194,7 +3208,7 @@ static void *migration_thread(void *opaque)
rcu_register_thread();
object_ref(OBJECT(s));
- s->iteration_start_time = qemu_clock_get_ms(QEMU_CLOCK_REALTIME);
+ update_iteration_initial_status(s);
qemu_savevm_state_header(s->to_dst_file);
@@ -3259,8 +3273,7 @@ static void *migration_thread(void *opaque)
* the local variables. This is important to avoid
* breaking transferred_bytes and bandwidth calculation
*/
- s->iteration_start_time = qemu_clock_get_ms(QEMU_CLOCK_REALTIME);
- s->iteration_initial_bytes = 0;
+ update_iteration_initial_status(s);
}
current_time = qemu_clock_get_ms(QEMU_CLOCK_REALTIME);
diff --git a/migration/savevm.c b/migration/savevm.c
index 412768216c..1ac15301ad 100644
--- a/migration/savevm.c
+++ b/migration/savevm.c
@@ -1456,6 +1456,7 @@ static int qemu_savevm_state(QEMUFile *f, Error **errp)
}
migrate_init(ms);
+ memset(&ram_counters, 0, sizeof(ram_counters));
ms->to_dst_file = f;
qemu_mutex_unlock_iothread();
--
2.21.0
next prev parent reply other threads:[~2019-08-15 16:59 UTC|newest]
Thread overview: 35+ messages / expand[flat|nested] mbox.gz Atom feed top
2019-08-15 16:34 [Qemu-devel] [PULL 00/33] migration queue Dr. David Alan Gilbert (git)
2019-08-15 16:34 ` [Qemu-devel] [PULL 01/33] migration: Add error_desc for file channel errors Dr. David Alan Gilbert (git)
2019-08-15 16:34 ` [Qemu-devel] [PULL 02/33] hw/net: fix vmxnet3 live migration Dr. David Alan Gilbert (git)
2019-08-15 16:34 ` [Qemu-devel] [PULL 03/33] migration: consolidate time info into populate_time_info Dr. David Alan Gilbert (git)
2019-08-15 16:34 ` [Qemu-devel] [PULL 04/33] migration/postcopy: the valid condition is one less then end Dr. David Alan Gilbert (git)
2019-08-15 16:34 ` [Qemu-devel] [PULL 05/33] migration/postcopy: break the loop when there is no more page to discard Dr. David Alan Gilbert (git)
2019-08-15 16:34 ` [Qemu-devel] [PULL 06/33] migration/postcopy: discard_length must not be 0 Dr. David Alan Gilbert (git)
2019-08-15 16:34 ` [Qemu-devel] [PULL 07/33] migration/postcopy: reduce one operation to calculate fixup_start_addr Dr. David Alan Gilbert (git)
2019-08-15 16:34 ` [Qemu-devel] [PULL 08/33] migration/postcopy: do_fixup is true when host_offset is non-zero Dr. David Alan Gilbert (git)
2019-08-15 16:34 ` [Qemu-devel] [PULL 09/33] migration/savevm: flush file for iterable_only case Dr. David Alan Gilbert (git)
2019-08-15 16:34 ` [Qemu-devel] [PULL 10/33] migration/savevm: split qemu_savevm_state_complete_precopy() into two parts Dr. David Alan Gilbert (git)
2019-08-15 16:34 ` [Qemu-devel] [PULL 11/33] migration/savevm: move non SaveStateEntry condition check out of iteration Dr. David Alan Gilbert (git)
2019-08-15 16:34 ` [Qemu-devel] [PULL 12/33] migration/postcopy: PostcopyState is already set in loadvm_postcopy_handle_advise() Dr. David Alan Gilbert (git)
2019-08-15 16:34 ` [Qemu-devel] [PULL 13/33] migration/postcopy: start_postcopy could be true only when migrate_postcopy() return true Dr. David Alan Gilbert (git)
2019-08-15 16:34 ` [Qemu-devel] [PULL 14/33] migration: use migration_in_postcopy() to check POSTCOPY_ACTIVE Dr. David Alan Gilbert (git)
2019-08-15 16:34 ` [Qemu-devel] [PULL 15/33] migration: just pass RAMBlock is enough Dr. David Alan Gilbert (git)
2019-08-15 16:34 ` [Qemu-devel] [PULL 16/33] migration: equation is more proper than and to check LOADVM_QUIT Dr. David Alan Gilbert (git)
2019-08-15 16:34 ` [Qemu-devel] [PULL 17/33] migration: return -EINVAL directly when version_id mismatch Dr. David Alan Gilbert (git)
2019-08-15 16:34 ` [Qemu-devel] [PULL 18/33] migration: extract ram_load_precopy Dr. David Alan Gilbert (git)
2019-08-15 16:34 ` [Qemu-devel] [PULL 19/33] migration/postcopy: make PostcopyDiscardState a static variable Dr. David Alan Gilbert (git)
2019-08-15 16:34 ` [Qemu-devel] [PULL 20/33] migration/postcopy: simplify calculation of run_start and fixup_start_addr Dr. David Alan Gilbert (git)
2019-08-15 16:34 ` [Qemu-devel] [PULL 21/33] migration/postcopy: use QEMU_IS_ALIGNED to replace host_offset Dr. David Alan Gilbert (git)
2019-08-15 16:34 ` [Qemu-devel] [PULL 22/33] hmp: Remove migration capabilities from "info migrate" Dr. David Alan Gilbert (git)
2019-08-15 16:34 ` [Qemu-devel] [PULL 23/33] migration: remove unused field bytes_xfer Dr. David Alan Gilbert (git)
2019-08-15 16:34 ` Dr. David Alan Gilbert (git) [this message]
2019-08-15 16:34 ` [Qemu-devel] [PULL 25/33] migration: add qemu_file_update_transfer interface Dr. David Alan Gilbert (git)
2019-08-15 16:34 ` [Qemu-devel] [PULL 26/33] migration: add speed limit for multifd migration Dr. David Alan Gilbert (git)
2019-08-15 16:34 ` [Qemu-devel] [PULL 27/33] migration: update ram_counters for multifd sync packet Dr. David Alan Gilbert (git)
2019-08-15 16:34 ` [Qemu-devel] [PULL 28/33] migration: rename migration_bitmap_sync_range to ramblock_sync_dirty_bitmap Dr. David Alan Gilbert (git)
2019-08-15 16:35 ` [Qemu-devel] [PULL 29/33] migration/postcopy: use mis->bh instead of allocating a QEMUBH Dr. David Alan Gilbert (git)
2019-08-15 16:35 ` [Qemu-devel] [PULL 30/33] qemu-file: move qemu_{get, put}_counted_string() declarations Dr. David Alan Gilbert (git)
2019-08-15 16:35 ` [Qemu-devel] [PULL 31/33] migration: Add traces for multifd terminate threads Dr. David Alan Gilbert (git)
2019-08-15 16:35 ` [Qemu-devel] [PULL 32/33] migration: Make global sem_sync semaphore by channel Dr. David Alan Gilbert (git)
2019-08-15 16:35 ` [Qemu-devel] [PULL 33/33] migration: add some multifd traces Dr. David Alan Gilbert (git)
2019-08-16 12:58 ` [Qemu-devel] [PULL 00/33] migration queue Peter Maydell
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20190815163504.18937-25-dgilbert@redhat.com \
--to=dgilbert@redhat.com \
--cc=ivanren@tencent.com \
--cc=marcandre.lureau@redhat.com \
--cc=qemu-devel@nongnu.org \
--cc=quintela@redhat.com \
--cc=richardw.yang@linux.intel.com \
--cc=yury-kotov@yandex-team.ru \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).