Problem mit Umlauten

Problem with Umlauts

You may find that files or folders with umlauts or other special characters in the names are not backed up. Then, the coding of the data that is created during the backup does not agree with the coding in the original file. Errors of this type are listed in the backup error logs file (dsmerrlog). 

Example for an extract of dsmerror.log-file:

To uncover an above-mentioned Error, you can find out the type of encoding as follows: 

Save the file name of the folder in question (in the example /home/test) recursively in a file (in the example dateinamen.out). With the command "file dateinamen.out" you can obtain information on the file type and thus for coding.

If the output is

• • x: ASCII text   => no special characters, no problem, the place for the backup locale doesn’t really matter.
  • x: UTF-8 Unicode text => all files are UTF-8 encoded. For backup, a UTF-8 locale will be set, which is standard on most modern operating systems.
  • x: Non-ISO extended-ASCII text or ISO-8859 text
    in the folder exist filenames that have not been created using an ISO-coded code page (e.g. Windows files, CP850, etc.) Simplest solution: change the file name in agreement with the users. Other solution: Try adjusting the following locales before starting the backup software (in our tests, this works):

    root@username-PC~$  export LANG="de_DE" root@username-PC~$  export LC_CTYPE=de_DE@euro root@username-PC~$  locale

    The output of the "locale" command displays the following:

    LANG=de_DE LC_CTYPE=de_DE@euro LC_NUMERIC="de_DE" LC_TIME="de_DE" LC_COLLATE="de_DE" LC_MONETARY="de_DE" LC_MESSAGES="de_DE" LC_PAPER="de_DE" LC_NAME="de_DE" LC_ADDRESS="de_DE" LC_TELEPHONE="de_DE" LC_MEASUREMENT="de_DE" LC_IDENTIFICATION="de_DE" LC_ALL=

IBM also provides detailed documents on the subject of backup files with special characters:

• TSM Client and UTF-8-coding (english)
• TSM Client and local setting(english)