[Bug 12835] New: Allow --link-dest to link to an optionally unexisting directory

[just subscribed for rsync-qa from bugzilla via rsync]

https://bugzilla.samba.org/show_bug.cgi?id=12835

Bug ID: 12835
   Summary: Allow --link-dest to link to an optionally unexisting
directory
   Product: rsync
   Version: 3.1.3
  Hardware: All
OS: All
Status: NEW
  Severity: enhancement
  Priority: P5
 Component: core
  Assignee: way...@samba.org
  Reporter: julien+bugzilla-sa...@palard.fr
QA Contact: rsync...@samba.org

Found myself trying to hardlink between versions, but as I may rsync to a new
server I may not having an old version to point to, so I had to use:

  rsync -rl --link-dest=$(ssh prod "cd /var/www; readlink -e current || echo
.") static index.html prod:/var/www/$CI_COMMIT_SHA/
  ssh prod "ln -nfs $CI_COMMIT_SHA" current 

Which is ugly.

I'd prefer something like:

  rsync -rl --link-dest=current static index.html prod:/var/www/$CI_COMMIT_SHA/
  ssh prod "ln -nfs $CI_COMMIT_SHA" current

But warning / erroring on missing --link-dest sound legitimate, and we may not
want to change this. If not, why not introduce a --link-dest-if-exists? Or a
--ignore-missing-link-dest (in combination with the normal --link-dest)?

-- 
You are receiving this mail because:
You are the QA Contact for the bug.

-- 
Please use reply-all for most replies to avoid omitting the mailing list.
To unsubscribe or change options: https://lists.samba.org/mailman/listinfo/rsync
Before posting, read: http://www.catb.org/~esr/faqs/smart-questions.html

[Bug 12806] Deleting in a row of hardlinked snapshots resets file permissions.

[just subscribed for rsync-qa from bugzilla via rsync]

https://bugzilla.samba.org/show_bug.cgi?id=12806

--- Comment #5 from Heinz-Willi Wyes  ---
Lars Ellenberg provided a workaround for the behaviour. Using

rsync -d --delete --super `mktemp -d`/ snapshots/2017-05-26-15-00-53/

plays the trick.

Not sure whether the --super option has implications not suitable for other
scenarios. For me it works just fine. Nevertheless, there should be a patch for
rsync in order to overcome the questionable behaviour.

-- 
You are receiving this mail because:
You are the QA Contact for the bug.

-- 
Please use reply-all for most replies to avoid omitting the mailing list.
To unsubscribe or change options: https://lists.samba.org/mailman/listinfo/rsync
Before posting, read: http://www.catb.org/~esr/faqs/smart-questions.html

Re: Bug: rsync erroneously changes modification time

[Paul Slootman via rsync]

On Mon 12 Jun 2017, max.power--- via rsync wrote:

> How exactly does rsync determine that the copy has the incorrect timestamp
> and not the source file?

The source by definition is correct.


Paul

-- 
Please use reply-all for most replies to avoid omitting the mailing list.
To unsubscribe or change options: https://lists.samba.org/mailman/listinfo/rsync
Before posting, read: http://www.catb.org/~esr/faqs/smart-questions.html

Re: Bug: rsync erroneously changes modification time

[Fabian Cenedese via rsync]

At 08:11 12.06.2017, max.power--- via rsync wrote:
>Content-Transfer-Encoding: 7bit
>Content-Disposition: inline
>
>How exactly does rsync determine that the copy has the incorrect  
>timestamp and not the source file?
>Does it assume that the copy must be incorrect or are there other  
>criteria that have to be considered?

rsync is only synching one way, it's always going from source to dest.
Therefore source is always 'correct' whereas dest will be adjusted to
match source. If you want a two way synching you'd need to call
rsync twice with different parameters (and switched source/dest)
or use a different tool.

bye  Fabi


-- 
Please use reply-all for most replies to avoid omitting the mailing list.
To unsubscribe or change options: https://lists.samba.org/mailman/listinfo/rsync
Before posting, read: http://www.catb.org/~esr/faqs/smart-questions.html

Re: Bug: rsync erroneously changes modification time

[max.power--- via rsync]

How exactly does rsync determine that the copy has the incorrect  
timestamp and not the source file?
Does it assume that the copy must be incorrect or are there other  
criteria that have to be considered?



Quoting Kevin Korb via rsync :


Whenever you use --times (included in --archive) rsync will fix
incorrect time stamps.  The only thing --size-only is doing is keeping
the incorrect data instead of replacing it.

The purpose of these options is to "fix" a copy done in a way that did
not preserve timestamps but the data is known to have not changed.
These options allow rsync to correct the incorrect timestamps without
even looking inside of the files.  If you are not 100% sure your file
data matches you should not be using --size-only.

On 06/11/2017 09:28 AM, max.power--- via rsync wrote:

When a file of same length already exists at the destination then the
command 'rsync --archive --size-only' (--archive is same as -rlptgoD)
may change the modification time of the destination file even if no
modification was made.

Type the following commands in a terminal in order to reproduce the
problem:

$ mkdir source
$ mkdir target
$ echo "file one" > source/file
$ echo "file two" > target/file
$ srcstamp=2011.11
$ tgtstamp=2101.00
$ touch -t $srcstamp source/file
$ touch -t $tgtstamp target/file
$ stat -c "%y %s" source/file
$ stat -c "%y %s" target/file

Notice, that although the file in source/ and target/ folder contain
different content they do not  differ in size. Thus, there following
command should not make any changes:

$ rsync -rlptgoD --size-only source/ target

There was indeed no backup performed, which can be verified by viewing
the contents of the files:

$ cat source/file
$ cat target/file

However, the modification time of the target file was updated to the
same modification time as file in source/:

$ stat -c "%y %s" source/file
$ stat -c "%y %s" target/file

Since the file was not modified the modification time should not be
changed. Omitting the '-t' option in above rsync does not solve the
problem because when a backup is actually performed then the
modification time will simply be set to the current system time instead
of the modification time of the source file.

The same behaviour can be observed when using the flags -rlptgoD (or
--archive) with the --checksum flag. Although it is highly unlikely that
the files will differ if the checksums are the same, the modficaition
should still not be changed for consistency reasons.




-

ONLY AT VFEmail! - Use our Metadata Mitigator to keep your email out of
the NSA's hands!
$24.95 ONETIME Lifetime accounts with Privacy Features! 15GB disk! No
bandwidth quotas!
Commercial and Bulk Mail Options!


--
~*-,._.,-*~'`^`'~*-,._.,-*~'`^`'~*-,._.,-*~'`^`'~*-,._.,-*~'`^`'~*-,._.,
Kevin Korb  Phone:(407) 252-6853
Systems Administrator   Internet:
FutureQuest, Inc.   ke...@futurequest.net  (work)
Orlando, Floridak...@sanitarium.net (personal)
Web page:   http://www.sanitarium.net/
PGP public key available on web site.
~*-,._.,-*~'`^`'~*-,._.,-*~'`^`'~*-,._.,-*~'`^`'~*-,._.,-*~'`^`'~*-,._.,





-

ONLY AT VFEmail! - Use our Metadata Mitigator to keep your email out of the 
NSA's hands!
$24.95 ONETIME Lifetime accounts with Privacy Features!  
15GB disk! No bandwidth quotas!
Commercial and Bulk Mail Options!  


--
Please use reply-all for most replies to avoid omitting the mailing list.
To unsubscribe or change options: https://lists.samba.org/mailman/listinfo/rsync
Before posting, read: http://www.catb.org/~esr/faqs/smart-questions.html