Re: [Bug-tar] incremental backups and --one-file-system (version 1.25)

[gene heskett]

On Monday, November 22, 2010 06:11:01 am Paul Eggert did opine:

> On 11/21/2010 10:40 PM, gene heskett wrote:
> > Aren't the modes amanda uses for incremental ever tested?
> 
> Apparently not.  Could you contribute some test cases along those lines?
> That would help prevent this sort of problem in the future.

I don't think anything more than a case for the --listed-incremental (I 
think that's the option that is failing), but let me see if an amgtar log 
shows it, I'll attach last nights failed run's log if it does.

Here is last nights summary section from that email:

(shop has an older tar, so doesn't suffer)

DUMP SUMMARY:
                                                 DUMPER STATS              
TAPER STATS
HOSTNAM DISK                        L ORIG-kB  OUT-kB COMP% MMM:SS    KB/s 
MMM:SS    KB/s
------------------------------------- ----------------------------------- 
-------------
coyote  /GenesAmandaHelper-0.6      2  194100   99214  51.1  0:31  3166.0  0:00 
992140.0
coyote  /bin                        0      10      10    --  0:00   110.4  0:00 
  100.0
coyote  /boot                       2   24900   24900    --  0:01 47068.9  0:01 
24900.0
coyote  /etc                        0      10       1  10.0  0:00    18.5  0:00 
    0.0
coyote  /home                       1 3575160 2111370  59.1  7:23  4770.3  0:59 
35785.9
coyote  /lib                        1  158150   52284  33.1  1:16   691.5  0:00 
522840.0
coyote  /opt                        1  194370  186699  96.1  0:32  5764.9  0:02 
93349.5
coyote  /root                       0      10       1  10.0  0:00    14.3  0:00 
    0.0
coyote  /sbin                       0      10      10    --  0:00    61.6  0:00 
  100.0
coyote  /tmp                        1   24870   20696  83.2  0:04  5880.0  0:00 
206960.0
coyote  /usr/X11R6                  0      10       1  10.0  0:00    22.8  0:00 
    0.0
coyote  /usr/bin                    1  434050  434050    --  0:59  7375.0  0:05 
86810.0
coyote  /usr/dlds/misc              0      10      10    --  0:00   208.9  0:00 
  100.0
coyote  /usr/dlds/rpms              0      10      10    --  0:00   109.3  0:00 
  100.0
coyote  /usr/dlds/tgzs              0      10      10    --  0:00   201.8  0:00 
  100.0
coyote  /usr/include                0      10       1  10.0  0:00    11.4  0:00 
    0.0
coyote  /usr/lib                    2  129070   41393  32.1  0:42   992.0  0:01 
41393.0
coyote  /usr/libexec                0      10       1  10.0  0:00    15.2  0:00 
    0.0
coyote  /usr/local                  2   37660    8963  23.8  0:08  1097.9  0:00 
89630.0
coyote  /usr/movies                 1 1437780 1437780    --  0:41 35416.2  0:22 
65353.6
coyote  /usr/music                  0      10      10    --  0:00    52.0  0:19 
    0.5
coyote  /usr/pix                    0      10      10    --  0:00   111.1  0:00 
  100.0
coyote  /usr/sbin                   0      10       1  10.0  0:00    10.9  0:00 
    0.0
coyote  /usr/share                  2   41580    8904  21.4  0:37   238.7  0:00 
89040.0
coyote  /usr/src                    2   46850    3359   7.2  2:12    25.5  0:00 
33590.0
coyote  /var                        1  421660  152829  36.2  1:45  1457.7  0:01 
152829.0
shop    /etc                        1     300      20   6.7  0:00   148.5  0:00 
  200.0
shop    /home                       1     200      10   5.0  0:00   100.2  0:00 
  100.0
shop    /usr/lib/amanda             0    3470    1113  32.1  0:02   490.7  0:00 
11130.0
shop    /usr/local                  1      40       1   2.5  0:05     0.2  0:00 
    0.0
shop    /usr/src                    0   55230    9911  17.9  0:18   546.1  0:00 
99110.0
shop    /var/lib/amanda             0    1500     314  20.9  0:00   855.4  0:00 
 3140.0

Ok, get the amgtar logs for coyote's /usr/sbin, which has
address@hidden Daily]# du -h /usr/sbin
61M     /usr/sbin
data in it.

Ok, so two files, the lower numbered one from the estimate request which 
is correct, the higher numbered one from the backup run itself which failed.

This, for us amanda users, is serious enough that both 1.24 and 1.25 
should be removed from the dl directories at gnu.org, to be replaced by 
1.26 when its fixed. 1.23 works well however, so it can remain visible.

You can send me a precompiled 32 bit x86 binary for testing, but while I 
would much rather keep it in the package system, I am not allergic to 
building the rpms from the tarball if your supplied scripts can do that.  
TBT I haven't looked and haven't built a tar locally since 1.13.
Yonks ago.  I do build amanda at least weekly from the tarballs, so the 
package system knows nothing about amanda.  I also build my own kernels.
In other words I know just enough to be dangerous. ;-)

Thanks Paul.

-- 
Cheers, Gene
"There are four boxes to be used in defense of liberty:
 soap, ballot, jury, and ammo. Please use in that order."
-Ed Howdershelt (Author)
You climb to reach the summit, but once there, discover that all roads
lead down.
                -- Stanislaw Lem, "The Cyberiad"

amgtar.20101122005201000.debug
Description: Text document

amgtar.20101122010805000.debug
Description: Text document