1 |
Simon Boulet <simon@××××.ca> writes: |
2 |
|
3 |
> Hi, |
4 |
> |
5 |
> This is a known issue with newer versions of GNU cp and Rsnapshot. |
6 |
> |
7 |
> http://sourceforge.net/mailarchive/forum.php?thread_id=8988105&forum_id=41320 |
8 |
> |
9 |
> Update to Rsnapshot 1.2.3. |
10 |
|
11 |
Excellent info... thanks. |
12 |
|
13 |
It seems there should be some safeguard in place to prevent a change |
14 |
like this from allowing rsnapshot to eat its previously created |
15 |
backups. |
16 |
|
17 |
After installing the new cp, rsnapshot ate one full hourly.* every 4 |
18 |
hours until only hourly.0 remains by rming hourly.5 and moving |
19 |
hourly.* to hourly.5. And of course it won't copy hourly.0 to |
20 |
hourly.1 so hourly.0 isn't even renewed either. Within a few hours |
21 |
your relied upon backups have chewed themselves into oblivion. |
22 |
|
23 |
I've scripted something that writes a .SUCCESS file to my rsnapshot |
24 |
base each run so if it isn't found right before running rsnapshot, |
25 |
then rsnapshot isn't run. |
26 |
|
27 |
I'm surprised there wasn't some similar safeguard built into |
28 |
rsnapshot. |
29 |
|
30 |
-- |
31 |
gentoo-user@g.o mailing list |