On 2021-07-30 09:05, Marcel Telka wrote:
On Fri, Jul 30, 2021 at 12:02:13PM -0300, Till Wegmueller wrote:
I assume better means it can detect create a diff faster?

No, this is not about the speed (primarily).  Just about the quality of
data published via the protocol.  For example, with rsync I can be sure
that timestamps are okay, for example.  With ftp it usually works too,
but http is not so good and there is possibility that I might end up
re-downloading the same content again and again just because http might
fail to advertise mtime change (or no change, to be precise).
I can't help but add, that rsync is far more efficient overall. As it is
only interested in differences (as already mentioned). Which results in
less bytes transferred on both sides. Are the OI servers paying for bytes
transferred (have a cap/limit)? This will also leave more bandwidth available
for all.
OK I don't run or manage the OI servers. But couldn't help myself. ;-)

--Chris

Over the years I switched all my mirrors to rsync from initial ftp (or
http) just because of this.

I can see the point. I also think a fast diff for mirroring is nice. I'll

Thanks.  Please let me know once there is rsync back.

have a look in the future to see to get that going. For now please switch to
HTTP(S)

Sure, will do.


Thank you.

--
~10yrs a FreeBSD maintainer of ~160 ports
~40yrs of UNIX

Attachment: 0xBDE49540.asc
Description: application/pgp-keys

_______________________________________________
oi-dev mailing list
[email protected]
https://openindiana.org/mailman/listinfo/oi-dev

Reply via email to