rsync - May 2023 - Per file Log output to understand hard-link decision

Hello,

I know it is a often discussed topic how rsync decide about using
hardlinks or copy a file. Even if content is unchanged problems are
often file permissions and owner ships. I know that.

Is it possible to configure rsync that way that it logs for each file
its decision about using a hardlink and if not why exactly it doesn't?

The background of my question:
I'm part of maintainer team of "Back In Time" a desktop backup
software
using rsync in the back. For years we have users reporting about the
hardlink problem. And we are sure that it isn't their fault but ours.
We can't reproduce the problems for sure but we observe the behavior
also on our own machines sometimes.

It would help our investigation if we could better understand the
hardlink-decision for each file and folder.

Kind
Christian

The only way I know of to determine this behavior is to use the 
--link-dest option in rsync OR use the much older cp -al then rsync over 
top of it method.  With --link-dest a change in the file's metadata 
causes rsync to duplicate the file in order to store both versions of 
the metadata.  With the old cp -al method rsync just makes the metadata 
change on the hard link already in the target which causes all instances 
of the file to be changed rather than the file being duplicated.

On 5/14/23 08:38, c.buhtz--- via rsync wrote:
> Hello,
> 
> I know it is a often discussed topic how rsync decide about using
> hardlinks or copy a file. Even if content is unchanged problems are
> often file permissions and owner ships. I know that.
> 
> Is it possible to configure rsync that way that it logs for each file
> its decision about using a hardlink and if not why exactly it doesn't?
> 
> The background of my question:
> I'm part of maintainer team of "Back In Time" a desktop
backup software
> using rsync in the back. For years we have users reporting about the
> hardlink problem. And we are sure that it isn't their fault but ours.
> We can't reproduce the problems for sure but we observe the behavior
> also on our own machines sometimes.
> 
> It would help our investigation if we could better understand the
> hardlink-decision for each file and folder.
> 
> Kind
> Christian
> 
-- 
~*-,._.,-*~'`^`'~*-,._.,-*~'`^`'~*-,._.,-*~'`^`'~*-,._.,-*~'`^`'~*-,._.,
	Kevin Korb			Phone:    (407) 252-6853
	Systems Administrator		Internet:
	FutureQuest, Inc.		Kevin at FutureQuest.net  (work)
	Orlando, Florida		kmk at sanitarium.net (personal)
	Web page:			https://sanitarium.net/
	PGP public key available on web site.
~*-,._.,-*~'`^`'~*-,._.,-*~'`^`'~*-,._.,-*~'`^`'~*-,._.,-*~'`^`'~*-,._.,

Don't know if this is enough for you, but it may help at least a bit to hunt
down your problem. There is a flag -i

 From man rsync
--itemize-changes, -i    output a change-summary for all updates

this gives either a "." for no change, or a letter code for each
change in "data, time, size" etc., for any filename it logs.

You will figure most flags. Others you may need to look up somewhere.

Hope this helps (a bit).

Hardy


Am 14.05.23 um 14:38 schrieb c.buhtz--- via rsync:
> Hello,
> 
> I know it is a often discussed topic how rsync decide about using
> hardlinks or copy a file. Even if content is unchanged problems are
> often file permissions and owner ships. I know that.
> 
> Is it possible to configure rsync that way that it logs for each file
> its decision about using a hardlink and if not why exactly it doesn't?
> 
> The background of my question:
> I'm part of maintainer team of "Back In Time" a desktop
backup software
> using rsync in the back. For years we have users reporting about the
> hardlink problem. And we are sure that it isn't their fault but ours.
> We can't reproduce the problems for sure but we observe the behavior
> also on our own machines sometimes.
> 
> It would help our investigation if we could better understand the
> hardlink-decision for each file and folder.
> 
> Kind
> Christian
>