ANS4042E: How to backup utf8 and iso-8859-15 files?

[c!w]

Hi,
During backup, i got a lot of failed objects.
We have servers with mixed charsets. Some webapps support utf8, other doesn't.
The trick with "export LANG=de_CH.iso88591" does not work. When I ran the scheduler with this command, all utf8-files are ignored!

Is there any trick to backup all kind of files, regardless of the charset?

Chris

 

[booman55]

You need to make sure the client version you're using supports UTF-8 characters. Then you'll need to set autofsrename=yes in client node configuration. This will create new filespaces that are unicode enabled the next time the backup runs.

 

[c!w]

You need to make sure the client version you're using supports UTF-8 characters. Then you'll need to set autofsrename=yes in client node configuration. This will create new filespaces that are unicode enabled the next time the backup runs.

Mhh, that did not work..

dsm.sys

SErvername  tsmsrv
   COMMmethod         TCPip
   TCPPort            1500
   TCPServeraddress   bar.foo.com

   AUTOFSRENAME       YES

I restarted the scheduler. But I still got this error codes.

10/08/08   02:14:12 ANS1228E Sending of object '/data/moodledata/18/Support/LernzieleAktivit<84>ten.doc' failed
10/08/08   02:14:12 ANS4042E Object name '/data/moodledata/18/Support/LernzieleAktivit<84>ten.doc' contains one or mor
e unrecognised characters and is not valid.

My tsm-client runs unicode enabled. All unicode files are going to be backuped, but all the iso encoded files are ignored.

cheers,
Chris

 

[vutcovici]

I had this problem in Ubuntu Linux.

Check that the en_US locale is installed
$ locale -a|grep en_US$
Generate the locale
$ sudo locale-gen en_US
$ locale -a|grep en_US$
en_US
# cd /opt/tivoli/tsm/client/ba/bin
# LC_ALL=en_US LANG=en_US ./dsmc incremental filename-with-problems.txt
# LC_ALL=en_US LANG=en_US ./dsmj
in scripts you can use: export LC_ALL=en_US LANG=en_US
References:
http://www.oucs.ox.ac.uk/hfs/help/faq/index.xml.ID=locale
http://www.adsm.org/lists/html/ADSM-L/2008-04/msg00207.html
http://lists.ibiblio.org/pipermail/unclug/2007-August/000404.htm

 

[Efim]

Hi
This is very good guide http://www-01.ibm.com/support/docview.wss?uid=swg27013395

Efim

 

[xyzegg]

I had this problem in Ubuntu Linux.

Check that the en_US locale is installed
$ locale -a|grep en_US$
Generate the locale
$ sudo locale-gen en_US
$ locale -a|grep en_US$
en_US
# cd /opt/tivoli/tsm/client/ba/bin
# LC_ALL=en_US LANG=en_US ./dsmc incremental filename-with-problems.txt
# LC_ALL=en_US LANG=en_US ./dsmj
in scripts you can use: export LC_ALL=en_US LANG=en_US
References:
http://www.oucs.ox.ac.uk/hfs/help/faq/index.xml.ID=locale
http://www.adsm.org/lists/html/ADSM-L/2008-04/msg00207.html
http://lists.ibiblio.org/pipermail/unclug/2007-August/000404.htm

Thank you! Here in Debian GNU/Linux 6.0, we do a little different than you, but your script help us to solve the problem. We just uncommented the line "en_US ISO-8859-1" and run locale-gen. After that, we could see en_US in the output of locale -a | grep en_US.

Kill the scheduler and dsmcad and start them again with LC_ALL=en_US LANG=en_US.

/etc/rc.local:

env LC_ALL=en_US LANG=en_US LANGUAGE=en_US nohup dsmc schedule > /dev/null 2>&1 &

env LC_ALL=en_US LANG=en_US LANGUAGE=en_US dsmcad > /dev/null 2>&1


We do a Onetime Incr and no more errors.


tsm: XPTO>q actlog node=XXX msg=4959

Date/Time Message
-------------------- ----------------------------------------------------------
07/04/2018 11:30:19 ANE4959I (Session: 64995, Node: XXX) Total number of
objects failed: 0 (SESSION: 64995)

tsm: XPTO>