qemu/qapi
Peter Xu 8b2395970a migration: Allow user to specify available switchover bandwidth
Migration bandwidth is a very important value to live migration.  It's
because it's one of the major factors that we'll make decision on when to
switchover to destination in a precopy process.

This value is currently estimated by QEMU during the whole live migration
process by monitoring how fast we were sending the data.  This can be the
most accurate bandwidth if in the ideal world, where we're always feeding
unlimited data to the migration channel, and then it'll be limited to the
bandwidth that is available.

However in reality it may be very different, e.g., over a 10Gbps network we
can see query-migrate showing migration bandwidth of only a few tens of
MB/s just because there are plenty of other things the migration thread
might be doing.  For example, the migration thread can be busy scanning
zero pages, or it can be fetching dirty bitmap from other external dirty
sources (like vhost or KVM).  It means we may not be pushing data as much
as possible to migration channel, so the bandwidth estimated from "how many
data we sent in the channel" can be dramatically inaccurate sometimes.

With that, the decision to switchover will be affected, by assuming that we
may not be able to switchover at all with such a low bandwidth, but in
reality we can.

The migration may not even converge at all with the downtime specified,
with that wrong estimation of bandwidth, keeping iterations forever with a
low estimation of bandwidth.

The issue is QEMU itself may not be able to avoid those uncertainties on
measuing the real "available migration bandwidth".  At least not something
I can think of so far.

One way to fix this is when the user is fully aware of the available
bandwidth, then we can allow the user to help providing an accurate value.

For example, if the user has a dedicated channel of 10Gbps for migration
for this specific VM, the user can specify this bandwidth so QEMU can
always do the calculation based on this fact, trusting the user as long as
specified.  It may not be the exact bandwidth when switching over (in which
case qemu will push migration data as fast as possible), but much better
than QEMU trying to wildly guess, especially when very wrong.

A new parameter "avail-switchover-bandwidth" is introduced just for this.
So when the user specified this parameter, instead of trusting the
estimated value from QEMU itself (based on the QEMUFile send speed), it
trusts the user more by using this value to decide when to switchover,
assuming that we'll have such bandwidth available then.

Note that specifying this value will not throttle the bandwidth for
switchover yet, so QEMU will always use the full bandwidth possible for
sending switchover data, assuming that should always be the most important
way to use the network at that time.

This can resolve issues like "unconvergence migration" which is caused by
hilarious low "migration bandwidth" detected for whatever reason.

Reported-by: Zhiyi Guo <zhguo@redhat.com>
Reviewed-by: Joao Martins <joao.m.martins@oracle.com>
Reviewed-by: Juan Quintela <quintela@redhat.com>
Signed-off-by: Peter Xu <peterx@redhat.com>
Signed-off-by: Juan Quintela <quintela@redhat.com>
Message-ID: <20231010221922.40638-1-peterx@redhat.com>
2023-10-17 09:14:32 +02:00
..
acpi.json qapi: Reformat doc comments to conform to current conventions 2023-05-10 10:01:01 +02:00
audio.json audio/pw: Pipewire->PipeWire case fix for user-visible text 2023-07-17 15:22:56 +04:00
authz.json qapi: Reformat doc comments to conform to current conventions 2023-05-10 10:01:01 +02:00
block-core.json qemu-img: map: report compressed data blocks 2023-09-20 17:46:01 +02:00
block-export.json qapi: Reformat doc comments to conform to current conventions 2023-05-10 10:01:01 +02:00
block.json qapi: Reformat recent doc comments to conform to current conventions 2023-07-26 14:51:36 +02:00
char.json ui/vc: do not parse VC-specific options in Spice and GTK 2023-09-04 14:57:37 +04:00
common.json qapi: Reformat doc comments to conform to current conventions 2023-05-10 10:01:01 +02:00
compat.json qapi: Reformat doc comments to conform to current conventions 2023-05-10 10:01:01 +02:00
control.json docs/interop: Convert qmp-spec.txt to rST 2023-05-22 10:21:01 +02:00
crypto.json qapi: Reformat doc comments to conform to current conventions 2023-05-10 10:01:01 +02:00
cryptodev.json spelling: information 2023-06-09 23:38:16 +03:00
cxl.json qapi: Reformat recent doc comments to conform to current conventions 2023-07-26 14:51:36 +02:00
dump.json qapi: Reformat doc comments to conform to current conventions 2023-05-10 10:01:01 +02:00
error.json qapi: Reformat doc comments to conform to current conventions 2023-05-10 10:01:01 +02:00
introspect.json qapi: Reformat doc comments to conform to current conventions 2023-05-10 10:01:01 +02:00
job.json qapi: Reformat doc comments to conform to current conventions 2023-05-10 10:01:01 +02:00
machine-target.json qapi: Reformat recent doc comments to conform to current conventions 2023-07-26 14:51:36 +02:00
machine.json accel/tcg: remove CONFIG_PROFILER 2023-06-26 17:33:00 +02:00
meson.build meson: Replace CONFIG_SOFTMMU -> CONFIG_SYSTEM_ONLY 2023-06-20 10:01:30 +02:00
migration.json migration: Allow user to specify available switchover bandwidth 2023-10-17 09:14:32 +02:00
misc-target.json qapi: Reformat doc comments to conform to current conventions 2023-05-10 10:01:01 +02:00
misc.json qapi: Correct "eg." to "e.g." in documentation 2023-07-25 17:20:32 +03:00
net.json net: add initial support for AF_XDP network backend 2023-09-18 14:36:13 +08:00
opts-visitor.c cutils: Adjust signature of parse_uint[_full] 2023-06-02 12:27:19 -05:00
pci.json qapi: Reformat doc comments to conform to current conventions 2023-05-10 10:01:01 +02:00
pragma.json qapi/pragma: Tidy up comments 2022-05-16 07:31:26 +02:00
qapi-clone-visitor.c qapi: Make visitor functions taking Error ** return bool, not void 2020-07-10 15:18:08 +02:00
qapi-dealloc-visitor.c qapi: Make visitor functions taking Error ** return bool, not void 2020-07-10 15:18:08 +02:00
qapi-forward-visitor.c qapi: remove needless include 2022-03-22 14:46:18 +04:00
qapi-schema.json docs/interop: Convert qmp-spec.txt to rST 2023-05-22 10:21:01 +02:00
qapi-type-helpers.c monitor: introduce HumanReadableText and HMP support 2021-11-02 15:55:13 +00:00
qapi-util.c qapi: Fix stale reference to scripts/qapi.py in a comment 2022-02-28 11:39:35 +01:00
qapi-visit-core.c qapi: Factor out compat_policy_input_ok() 2021-10-29 21:27:20 +02:00
qdev.json qapi/qdev: Tidy up device_add documentation 2023-07-26 14:51:36 +02:00
qmp-dispatch.c monitor: mark mixed functions that can suspend 2023-04-20 11:17:35 +02:00
qmp-event.c Replace qemu_gettimeofday() with g_get_real_time() 2022-04-06 10:50:37 +02:00
qmp-registry.c qapi: Generalize command policy checking 2021-10-29 18:24:46 +02:00
qobject-input-visitor.c include: add qemu/keyval.h 2022-04-21 17:03:51 +04:00
qobject-output-visitor.c qapi: Extend -compat to set policy for unstable interfaces 2021-10-29 21:28:01 +02:00
qom.json backends/hostmem-file: Add "rom" property to support VM templating with R/O files 2023-09-19 10:23:21 +02:00
rdma.json qapi: Reformat doc comments to conform to current conventions 2023-05-10 10:01:01 +02:00
replay.json qapi: Reformat doc comments to conform to current conventions 2023-05-10 10:01:01 +02:00
rocker.json qapi: Reformat doc comments to conform to current conventions 2023-05-10 10:01:01 +02:00
run-state.json qapi: Reformat doc comments to conform to current conventions 2023-05-10 10:01:01 +02:00
sockets.json qapi: Reformat doc comments to conform to current conventions 2023-05-10 10:01:01 +02:00
stats.json qapi: Reformat doc comments to conform to current conventions 2023-05-10 10:01:01 +02:00
string-input-visitor.c qapi, qemu-options: make all parsing visitors parse boolean options the same 2020-11-04 12:00:40 -05:00
string-output-visitor.c qapi: remove needless include 2022-03-22 14:46:18 +04:00
tpm.json qapi: Reformat doc comments to conform to current conventions 2023-05-10 10:01:01 +02:00
trace-events qapi: Generalize struct member policy checking 2021-10-29 18:23:09 +02:00
trace.h trace: switch position of headers to what Meson requires 2020-08-21 06:18:24 -04:00
trace.json qapi: Reformat recent doc comments to conform to current conventions 2023-07-26 14:51:36 +02:00
transaction.json qapi: Reformat doc comments to conform to current conventions 2023-05-10 10:01:01 +02:00
ui.json qapi: Reformat recent doc comments to conform to current conventions 2023-07-26 14:51:36 +02:00
virtio.json qapi: Reformat doc comments to conform to current conventions 2023-05-10 10:01:01 +02:00
yank.json qapi: Reformat doc comments to conform to current conventions 2023-05-10 10:01:01 +02:00