Hi Nivin,
On Wed, Mar 22, 2017 at 12:07:19PM -0700, Nivin Lawrence wrote:
> I was unable to find a lot of details on NBD over UDP, could you please
> let me know if this is supported (for read/write).
It isn't; NBD relies on the TCP sequencing and checksumming far too much
currently.
> The objective is to achieve
> the same functionality of server restart by keeping the server stateless so
> that if the server has to be restarted, you could still have the clients
> continue, which does not look like is the case with TCP. Possibly NBD has more
> reasons than just TCP to support server restart, if so, could you please list
> a
> few.
>
> If not UDP, is there an option to restart the NBD server without impacting
> the clients?
nbd-client has a "-persist" option which is meant to support connection
restart when necessary, but I'm not sure how well it works currently
(you'd have to check). I believe we fixed a bug where it would give up
on the first RST packet it received (which would mean it wouldn't work
for any type of server restart), but I'm not certain about that.
There are some kernel changes afoot which should help improve this
situation; they're not released yet, however.
--
< ron> I mean, the main *practical* problem with C++, is there's like a dozen
people in the world who think they really understand all of its rules,
and pretty much all of them are just lying to themselves too.
-- #debian-devel, OFTC, 2016-02-12
------------------------------------------------------------------------------
Check out the vibrant tech community on one of the world's most
engaging tech sites, Slashdot.org! http://sdm.link/slashdot
_______________________________________________
Nbd-general mailing list
Nbd-general@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/nbd-general
