On 2026-06-20 20:07, Andrea Venturoli wrote:
On 6/20/26 19:24, Daniel Engberg wrote:
Hi,
Thanks for answering.
To my knowledge it's the last version that's "fully" or close to fully
working and/or somewhat "validated" to work all others are considered
WIP .
I guess "validated" doesn't really have an "official" meaning here :)
Personally I've been using only 4.22/4.23 for a long time with
satisfaction. Of course my use cases can't possibly cover the myriad of
options, ways of using it, interactions with other softwares, etc...
I wonder, however, if these newer versions are "WIP" just because they
weren't tested by enough people (so eventual bugs were not reported and
not solved).
IMVVVVHO we should move forward and are very late at it.
Is there a list of what doesn't work in later versions?
Looking into bugzilla, I see #284603: that's against 4.20 however, which
is already obsolete. It's not clear if that's also bugging 4.22/4.23,
and, frankly, I doubt it works with 4.16 (I haven't tried it in a long
time, but I remember having had that problem in the past).
Then there's #281312, but it only lists a few ports that pinpoint
samba416, possibly just because that's the default. No indication of
whether that's an hard requirement or if someone tried them with a newer
version.
Personally I would be happy with ksmbd
Didn't even know about this.
I guess we won't see it anytime soon.
but that's not really portable at least in its current form.
<rant>
Samba is also less and less portable...
</rant>
:)
bye
av.
Hi,
As I've not followed this closely I may be very wrong on the subject so
please keep that in mind but I have a vague recollection of provisioning
(AD) being broken on ZFS and that earlier attempts didn't fix it at
least in beginning when trying to migrate to a newer version.
I'm not sure if that's still the case and at the time to my knowledge it
was the intended set of patches to be upstreamed but that seemingly
didn't occur (I'm not aware of the cause).
At least I got the impression of later iterations that they were
~testing/beta quality with the "maybe this will work" approach. Nothing
wrong with this at all and many thanks for everyone working on Samba but
this is probably the reason to why we're still stuck on 4.16 including
myself because in worst case you'll see (silent) data corruption which
is bad and possibly that provisioning is still broken.
We still carry a lot of patches in more recent versions, have there been
or are there any attempts of upstreaming?
To clarify, ksmbd apart from being tied to the Linux kernel focuses on
file transferring (and performance but I think that would be secondary
in our case at least initially), which probably is what most use it for
including myself without the other functionality.
Best regards,
Daniel
