qemu-block
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Qemu-block] [PATCH 3/3] qemu-iotests: Test postcopy migration


From: Kevin Wolf
Subject: Re: [Qemu-block] [PATCH 3/3] qemu-iotests: Test postcopy migration
Date: Wed, 19 Apr 2017 16:38:38 +0200
User-agent: Mutt/1.5.21 (2010-09-15)

Am 13.04.2017 um 20:44 hat Eric Blake geschrieben:
> On 04/13/2017 12:23 PM, Kevin Wolf wrote:
> > Signed-off-by: Kevin Wolf <address@hidden>
> > ---
> >  tests/qemu-iotests/181     | 117 
> > +++++++++++++++++++++++++++++++++++++++++++++
> >  tests/qemu-iotests/181.out |  38 +++++++++++++++
> >  tests/qemu-iotests/group   |   1 +
> >  3 files changed, 156 insertions(+)
> >  create mode 100755 tests/qemu-iotests/181
> >  create mode 100644 tests/qemu-iotests/181.out
> > 
> 
> I don't see a 180.out mentioned on-list anywhere (but the choice of test
> number is not too serious).

It's used by this one: [PATCH 00/18] nbd: BLOCK_STATUS

> ./check -raw 181
> 
>  === Write something on the source ===
> 
> +WARNING: Image format was not specified for 'TEST_DIR/t.raw' and
> probing guessed raw.
> +Automatically detecting the format is dangerous for raw images, write
> operations on block 0 will be restricted.
> +Specify the 'raw' format explicitly to remove the restrictions.
> 
> 
> You'll want to fix that.

Thanks for catching this, will fix.

> > +# Slow down migration so much that it definitely won't finish before we can
> > +# switch to postcopy
> > +silent=yes
> > +_send_qemu_cmd $src 'migrate_set_speed 4k' "(qemu)"
> > +_send_qemu_cmd $src 'migrate_set_capability postcopy-ram on' "(qemu)"
> > +_send_qemu_cmd $src "migrate -d unix:${MIG_SOCKET}" "(qemu)"
> > +_send_qemu_cmd $src 'migrate_start_postcopy' "(qemu)"
> > +
> > +QEMU_COMM_TIMEOUT=1 qemu_cmd_repeat=10 silent=yes \
> > +    _send_qemu_cmd $src "info migrate" "completed\|failed"
> > +silent=yes _send_qemu_cmd $src "" "(qemu)"
> 
> Do we want to tell the source to go back to a higher migrate_set_speed
> once we know we are in postcopy mode?  But the test passed for me under
> -qcow2 in less than 2 seconds, on an unloaded machine, so hopefully even
> on a heavily loaded machine we aren't giving up due to timeouts.

After switching to postcopy mode, the limits aren't applied any more
anyway (because the running VM is actively waiting for the pages and we
don't want to stall it artificially).

Kevin

Attachment: pgp3PNMIXne5e.pgp
Description: PGP signature


reply via email to

[Prev in Thread] Current Thread [Next in Thread]