Re: pkg/36656: amanda-client will fail to backup the requested filesystem

[Wolfgang Stukenbrock <Wolfgang.Stukenbrock%nagler-company.com@localhost>]

Hi - sorry for the delay ..

Setting the stat-buffers to zero will not solve the problem.
It just reduces the possibility that the problem is triggered!
It may happen still for all entries in /etc/fstab where stat() returns 
-1 as st_dev.

I've just checked that with 2.5.2p1.

The code in amanda is nonsence and broken - from my opinion.
There cannot be any reason to consult /etc/fstab and replace the path 
send from the server to be backed up!
So the best sollution is to eleminate this lookup.

The following patch disables the lookup.

*** client-src/getfsent.h.orig  Tue Jul 10 15:30:03 2007
--- client-src/getfsent.h       Tue Jul 10 15:31:22 2007
***************
*** 62,65 ****
--- 62,73 ----

  char *amname_to_fstype(char *str);

+ /*
+  * never look into fstab for filesystem directories - this is 
nonesence and will fail sometimes
+  * and results in backing up e.g. /proc instead ...
+  */
+ #ifndef IGNORE_FSTAB
+ #define IGNORE_FSTAB 1
+ #endif
+
  #endif /* ! GETFSENT_H */

W. Stukenbrock

Klaus Heinz wrote:

> The following reply was made to PR pkg/36656; it has been noted by GNATS.
>
> From: Klaus Heinz <k.heinz%oktsieben.kh-22.de@localhost>
> To: gnats-bugs%NetBSD.org@localhost
> Cc: 
> Subject: Re: pkg/36656: amanda-client will fail to backup the requested filesystem
> Date: Sat, 6 Oct 2007 02:19:48 +0200
>
>  Wolfgang Stukenbrock wrote:
>  
>  > 	remark: I've didn't send fix bug-report to the amanda developpers,
>  >           because I don't know a way to do it and there are lots of fixed
>  >           in the pkgsrc-tree for amanda. I think that one shoul be an
>  >           addirional fix to original distribution.
>  
>  You can file bugs at Sourceforge, see
>  
>    http://sourceforge.net/tracker/?group_id=120&atid=100120
>  
>  
>  A comparison between 2.4.4p4, 2.4.5p1 and 2.5.2p1 (latest stable version)
>  shows that the bug is probably fixed in the latter by using
>  
>    memset(stats, 0, SIZEOF(stats));
>  
>  on the structure array "stats".
>  
>  Personally, I would tend to do the following: Prepare a patch based on
>  the differences between 2.4.4p4 and 2.5.2p1 and ask on the amanda
>  mailing list for comments about your analysis of the problem and that patch.
>  If the resulting patch fixes the problem, pkgsrc could probably incorporate
>  this.
>  
>  ciao
>       Klaus
>    
>