[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [Qemu-devel] [PATCH] migration: regain control of images when migrat
From: |
Kevin Wolf |
Subject: |
Re: [Qemu-devel] [PATCH] migration: regain control of images when migration fails to complete |
Date: |
Wed, 18 May 2016 13:42:08 +0200 |
User-agent: |
Mutt/1.5.21 (2010-09-15) |
Am 18.05.2016 um 13:13 hat Dr. David Alan Gilbert geschrieben:
> * Greg Kurz (address@hidden) wrote:
> > We currently have an error path during migration that can cause
> > the source QEMU to abort:
>
> Hmm, wasn't there something similar recently, sorry I can't remember the
> details, but
> cc'ing some block people who might remember.
Not sure, but in any case calling bdrv_invalidate_cache_all() in the
error case sounds right to me.
I think I would have tried to write the patch a bit differently so that
it uses the normal error handling patterns, but that doesn't make it
wrong. Specifically, checking the error condition twice in two different
places seems strange; normally you check it once and then goto fail,
where the cleanup is done (which could include a call to
bdrv_invalidate_cache_all()).
Kevin
> Dave
> >
> > migration_thread()
> > migration_completion()
> > runstate_is_running() ----------------> true if guest is running
> > bdrv_inactivate_all() ----------------> inactivate images
> > qemu_savevm_state_complete_precopy()
> > ... qemu_fflush()
> > socket_writev_buffer() --------> error because destination fails
> > qemu_fflush() -------------------> set error on migration stream
> > migration_completion() -----------------> set migrate state to FAILED
> > migration_thread() -----------------------> break migration loop
> > vm_start() -----------------------------> restart guest with inactive
> > images
> >
> > and you get:
> >
> > qemu-system-ppc64: socket_writev_buffer: Got err=104 for
> > (32768/18446744073709551615)
> > qemu-system-ppc64:
> > /home/greg/Work/qemu/qemu-master/block/io.c:1342:bdrv_co_do_pwritev:
> > Assertion `!(bs->open_flags & 0x0800)' failed.
> > Aborted (core dumped)
> >
> > If we try postcopy with a similar scenario, we also get the writev error
> > message but QEMU leaves the guest paused because entered_postcopy is true.
> >
> > We could possibly do the same with precopy and leave the guest paused.
> > But since the historical default for migration errors is to restart the
> > source, this patch adds a call to bdrv_invalidate_cache_all() instead.
> >
> > Signed-off-by: Greg Kurz <address@hidden>
> > ---
> > migration/migration.c | 9 +++++++++
> > 1 file changed, 9 insertions(+)
> >
> > diff --git a/migration/migration.c b/migration/migration.c
> > index 991313a8629a..5726959ddfd9 100644
> > --- a/migration/migration.c
> > +++ b/migration/migration.c
> > @@ -1568,8 +1568,17 @@ static void migration_completion(MigrationState *s,
> > int current_active_state,
> > ret = bdrv_inactivate_all();
> > }
> > if (ret >= 0) {
> > + Error *local_err = NULL;
> > +
> > qemu_file_set_rate_limit(s->to_dst_file, INT64_MAX);
> > qemu_savevm_state_complete_precopy(s->to_dst_file, false);
> > +
> > + if (qemu_file_get_error(s->to_dst_file)) {
> > + bdrv_invalidate_cache_all(&local_err);
> > + if (local_err) {
> > + error_report_err(local_err);
> > + }
> > + }
> > }
> > }
> > qemu_mutex_unlock_iothread();
> >
> --
> Dr. David Alan Gilbert / address@hidden / Manchester, UK