Re: [Nbd] NBD_CMD_RESIZE
Hi all,
> Agreed.
>
>
> I'd be open to patches, if you have some. Doesn't appear it would take
> much in the kernel to get this functioning.
>
I'm glad there's interest in this, although I'm afraid I don't have
patches at the moment, and won't really have any time to work on it for
the next month or so. It's the best way to implement feature #3 on my
TODO list ^^.
We don't actually use the kernel client or canonical server (at the
moment :p); ideal from my point of view would be for the command +
semantics to be added to a protocol document, so we could write these
patches for the code we're most familiar with, and are actually going to
use, first. Still, it does feel like the spec *is* the canonical
implementation a lot of the time, so I could understand why that might
not be good enough from your point of view, so if adding the
functionality to nbd-client and nbd-server too is the cost of getting
the feature, it's one I'm fine with paying.
As for those semantics - what to do when multiple clients are connected
seems to be the only oddity. This is much more valuable to me (and
generally, I guess) if the client issuing the request can stay
connected, but there's a case that we "should" disconnect other clients,
since there's no way to inform them that the size has changed (and I'm
not hot on the idea of adding a mechanism to do so). Maybe make it a
command-line option ;).
/Nick
/Nick
Reply to: