Re: [Qemu-devel] [PATCH 9/9] block-migration: add named dirty bitmaps migration

[Paolo Bonzini]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1



On 08/01/2015 23:45, Eric Blake wrote:
>>> The bitmaps are transmitted many times in their entirety, but
>>> only the last copy actually means something.  The others are
>>> lost.  This means you should use the non-live interface
>>> (register_savevm).  This will simplify the code a lot.
> If I understand correctly, you are saying that:
> 
> block migration vs. assuming shared storage during migration
> 
> is independent from:
> 
> current behavior of resetting dirty tracking on destination vs.
> new one-shot non-live dirty bitmap migration
> 
> and that all four combinations are potentially useful.

Yes.  For example, you could use drive-mirror to snoop writes to an
NBD destination (e.g. to a malware scanner or to a secondary machine
if you have fault tolerance).  Then you need to migrate the dirty
bitmap even if you have shared storage.

> Also, by doing dirty bitmap migration as one-shot at the tail end
> of migration (in the small window when things are no longer live)
> you have a lot less effort, although the window of non-live
> activity is now a bit longer if the dirty table migration is not
> efficient.

As of this series, dirty bitmap migration is not attempting to track
the dirtiness of the dirty bitmap itself (you might have to read that
twice :)).  So there's nothing to lose in terms of efficiency.

Paolo
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v2

iQEcBAEBAgAGBQJUrwlwAAoJEL/70l94x66DSN4H+gLqqkghOqdkPzduTmf1hzi1
pwLe/TGZv+gNNWu8G2hO2FfasMlmBxEOVtwODFcwGExvPnH2jv1N7/dplIwKqbLS
g9Hq5cI3UagxfFQts7fyBhjCxeUrHuaKfIvc/A1kfMMwesn8PPGFUGEXNqIs1NGc
I+3qTE5TlcOKWCfL+3zgLDUowvRACbTJngHfveOtoWscVz743yQxhH3NOKPu7jxR
8H5AC9TK7sr3azHrXFgMP53W2qHT0KHTUyLQem+iKagFLsxX6oyOEn0ueMooAndE
tzvw+5EiFj8MfAzPAdQWEPYwxQyDBy/oXQizQTCmAQku2TZIlWi+kStJ3K/qEaM=
=v5Gj
-----END PGP SIGNATURE-----