Hello everyone
Thanks so much for the explanations Eric.
As a workaround, I will update the file daemon before the next full backup.
Márcio
Em 07/12/2021 05:52, Eric Bollengier via Bacula-users escreveu:
Hello Marcio,
On 12/6/21 19:24, Márcio (Departamento de Informática - PMI) wrote:
Hello everyone
Investigating a little more about what might be happening, I compared
the records (from the Bacula catalog in the database) of a file that
was backed up by 9.6.7 and 11.0.5, without having changed during
these two jobs .
I noticed that the "lstat" field of the "filename" table is
calculated differently between the two versions, although the file
hashes, stored in the "md5" field are the same, see:
jobid | pathid | filename | lstat
| md5
-------+--------+------------+------------------------------------------------------------+-----------------------------
14469 | 83108 | BPAGAR.DBF | A BAAAAAQCz IH/ B A A A Iu A A BOk0Kn
xxyqA BQdcaU A A M g | TYUkCmxYjmRvmCoTaOBDO+CRJx8
14749 | 83108 | BPAGAR.DBF | A A IH/ B A A A Iu A A BOk0Kn xxyqA
xxyqA A A M g | TYUkCmxYjmRvmCoTaOBDO+CRJx8
Now, why is this field calculated differently?
I'm not sure that part of the code changed a lot on Linux, however on
Windows it has been updated to use the snapshot to get file attributes
and handle some specific Windows files and corner cases. I don't know
if your filedaemon is a linux or a windows one.
The .bvfs_decode_lstat bconsole function can be used to decode this
field, and
we can see in your example a lot of "funny" values
*.bvfs_decode_lstat lstat="A BAAAAAQCz IH/ B A A A Iu A A BOk0Kn xxyqA
BQdcaU A A M g"
st_nlink=1
st_mode=33279
perm=-rwxrwxrwx
st_uid=0
st_gid=0
st_size=558
st_blocks=0
st_ino=281474976776371
st_ctime=1349895828
st_mtime=835136128
st_atime=1318273703
st_dev=0
LinkFI=0
*.bvfs_decode_lstat lstat="A A IH/ B A A A Iu A A BOk0Kn xxyqA xxyqA A
A M g"
st_nlink=1
st_mode=33279
perm=-rwxrwxrwx
st_uid=0
st_gid=0
st_size=558
st_blocks=0
st_ino=0
st_ctime=835136128
st_mtime=835136128
st_atime=1318273703
st_dev=0
LinkFI=0
One time the ino value is 0, and other time it's probably -1, the
mtime does not correspond to the ctime. You can get these attributes
with the shell command stat and compare with what Bacula has computed.
Hope it helps!
Best Regards,
Eric
_______________________________________________
Bacula-users mailing list
Bacula-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/bacula-users
--
_______________________________________________
Bacula-users mailing list
Bacula-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/bacula-users
