Hey folks.
I have a machine that I use as an intermediary to rsync between a NAS in
another building (the users there have poor bandwidth and need local
storage) and the our Netapp located in our Datacenter. I get lots of the
rsync: failed to set times on... errors, and files which have
I know from a lot of NAS boxes that they tend to use their internal
time to stamp files instead of the time given by a copy job.
The easiest way to test is to deliberately set the time off by a few
hours on the box you monted the stuff on, the NAS and netapp (or the
PS: Another (ugly) workaround: Use two linux boxes, place then both on
each side of the slow line. One side having the NAS mounted + running
rsync server, the other having the netapp mounted.
Then sync between those two linux boxes. Even if you have to use -c or
--ignore-times the full read
PPS. Yet another (a bit ugly) workaround, regarding how often the job is
done : try the options --modify-window=N (N in seconds), to allow timestamp
differences, or/and --size-only (comparing on filesize only).
Greg
On Mon, Apr 9, 2012 at 10:32 PM, Joachim Otahal (privat) j...@gmx.netwrote:
I
Thanks for the suggestions, guys.
I'll try testing times on the NAS device, as well as the --size-only
options. These are photos - I wonder what the odds are of a modified file
having the same size as the original?
Bill Dorrian
Network Administrator
Desk: 904-273-7625
Cell: 904-859-9471
Bill Dorrian wrote:
These are photos - I wonder what the odds are of a modified file
having the same size as the original?
If someone modifies the EXIF metadata (say, to correct a 'picture taken on'
timestamp for a camera that wasn't properly synchronized), the file size
would likely remain the
But the timestamp would not.
-Original Message-
From: rsync-boun...@lists.samba.org [mailto:rsync-boun...@lists.samba.org] On
Behalf Of Kyle Lanclos
Sent: Monday, April 09, 2012 5:06 PM
To: billdorr...@pgatourhq.com
Cc: rsync@lists.samba.org
Subject: Re: Problem syncing to Netapp
Stier, Matthew schrieb:
But the timestamp would not.
Be careful with that, I had cases where picture editors kept the
timestamps even if they did change the content. Only atime was changed,
mtime stayed. The affected users had the option selected in the program
(they thought of it as a