Marc Haber wrote:
> So, the ANF does seem to suppress the new .1.gz files from being
> reported as new, and the ARF does seem to suppress the removed .6.gz
> files from being reported as removed, but I don't understand what
> happens with the _not_ new .2.gz files (they come from mv .1.gz .2.gz)
> are reported as new, and why the _not_ removed .5.gz files (they go to
> mv .5.gz to .6.gz) are reported as removed.

What I think is happening, is that when the aide.db is created, a 
point-in-time snapshot (A) of your files is made:

inode   filename
10001   error.log
10002   error.log.0
10003   error.log.1.gz
10004   error.log.2.gz
10005   error.log.3.gz
10006   error.log.4.gz
10007   error.log.5.gz
10008   error.log.6.gz

Now, the next day when aide is run, error.log has become error.log.0, 
and error.log.1.gz is a new file. error.log.6.gz is removed. This looks 
like (B):

inode   filename
10010   error.log
10001   error.log.0
10011   error.log.1.gz
10003   error.log.2.gz
10004   error.log.3.gz
10005   error.log.4.gz
10006   error.log.5.gz
10007   error.log.6.gz

Again the next day, this will look like (C):

inode   filename
10020   error.log
10010   error.log.0
10021   error.log.1.gz
10011   error.log.2.gz
10003   error.log.3.gz
10004   error.log.4.gz
10005   error.log.5.gz
10006   error.log.6.gz

So when comparing C with A, error.log.2.gz is a new file and the 
original error.log.5.gz was removed.

I think for your ruleset to work, you need to update your aide.db every 
day (so at point B). So C gets compared to B and not A.

Sincerely,

Richard van den Berg
_______________________________________________
Aide mailing list
[email protected]
https://mailman.cs.tut.fi/mailman/listinfo/aide

Reply via email to