Page 1 of 2

Can't start DS server!

Posted: Tue Sep 13, 2005 7:08 am
by Luk
Hi!
I can't start DS derver on Unix. I'm getting following info:
Unable to open NLS install directory "".
Unable to get National Language Support shared memory segment.
Unable to get National Language Support shared memory segment.
Unable to get National Language Support shared memory segment.
/home/dsadm/Ascential/DataStage/DSEngine/sample/ds.rc: line 418: [: : integer expression expected
DataStage Engine 7.5.1.1 instance "ade" has been brought up.
After that dsrpc process is running but I can not connect from my client because UV internal error (39202)
Do you know what is going on? How can I fix it?

Posted: Tue Sep 13, 2005 7:10 am
by Sainath.Srinivasan
How did you check that the DS is up? Maybe someone was logged in when it was brought down.

Re: Can't start DS server!

Posted: Tue Sep 13, 2005 7:14 am
by PhilHibbs
Luk wrote:Hi!
I can't start DS derver on Unix. I'm getting following info:
Unable to open NLS install directory "".
Unable to get National Language Support shared memory segment.
Unable to get National Language Support shared memory segment.
Unable to get National Language Support shared memory segment.
/home/dsadm/Ascential/DataStage/DSEngine/sample/ds.rc: line 418: [: : integer expression expected
DataStage Engine 7.5.1.1 instance "ade" has been brought up.
What is in that ds.rc file at line 418?

Take a look in your DSEngine/dsenv file and see if that has become corrupted.

Posted: Tue Sep 13, 2005 7:16 am
by Luk
No! I am sure that at this moment only me is log in! I've checked it by

netstat -a

and *.dsrpc is LISTEN

Posted: Tue Sep 13, 2005 7:20 am
by Luk

Code: Select all

# start the cache daemon if configured
                val1=`bin/smat -t|grep DISKCACHE|tr -d " "|cut -d= -f2`
                val2=`bin/smat -t|grep DCWRITEDAEMON|tr -d " "|cut -d= -f2`
     ---->    if [ "$val1" -gt "0" -a "$val2" -gt "0" ]
                then
                        # cache daemon requested to start
                        bin/${DCFdaemon} START $val2 &
                fi

                echo "DataStage Engine $version instance \"$INSTANCETAG\" has been brought up."
line 418 is poited by ----->

Posted: Tue Sep 13, 2005 8:58 am
by PhilHibbs
Luk wrote:

Code: Select all

# start the cache daemon if configured
                val1=`bin/smat -t|grep DISKCACHE|tr -d " "|cut -d= -f2`
                val2=`bin/smat -t|grep DCWRITEDAEMON|tr -d " "|cut -d= -f2`
     ---->    if [ "$val1" -gt "0" -a "$val2" -gt "0" ]
                then
                        # cache daemon requested to start
                        bin/${DCFdaemon} START $val2 &
                fi

                echo "DataStage Engine $version instance "$INSTANCETAG" has been brought up."
line 418 is poited by ----->
OK when I run those two commands:

Code: Select all

bin/smat -t|grep DISKCACHE|tr -d " "|cut -d= -f2
bin/smat -t|grep DCWRITEDAEMON|tr -d " "|cut -d= -f2
the first returns -1, the second returns 0. Try them yourself while in the DSEngine directory in a shell prompt logged in as dsadm.

Have you checked your dsenv file?

Posted: Tue Sep 13, 2005 9:03 am
by Luk
yes, my dsenv is ok!
I have tried uvregen, but it doesn't work!

finally i reinstall DS server and now it is OK, but what was wrong and why, I still don't know.

If you have any ideas what was wrong and why please write it :)

regards

Posted: Tue Sep 13, 2005 9:10 am
by singhald
hi

Did u check the NLS INSTALLED in ur machine or not.
this is option u have to check while installing the DS

deepak singhal

Posted: Tue Sep 13, 2005 9:13 am
by Luk
yes NLS is installed

Posted: Tue Sep 13, 2005 9:23 am
by Luk
I must say that mu problem started when I was try to display very huge LOG (1,9GB) in Director and Directror stops responding. Is it possible that this caused my problem?

and I 've extra question: can I safely delete this file directory RT.LOG40?
Or how can I clear this log, but without using Director?

Posted: Tue Sep 13, 2005 9:32 am
by ArndW
Luk, go to the Administrator or to the DS command line and execute a "CLEAR.FILE RT_LOG40".

Posted: Tue Sep 13, 2005 9:46 am
by Luk
thanks Arnd! I used more dirty method: I have deleted manualy directory RT_LOG 40 and then I have manualy created directory RT_LOG40 :)
It is not very cleaver and betiful method but it works :)
next time I will do it through command line :)

Posted: Tue Sep 13, 2005 10:29 am
by ArndW
Luk,

I think that now your RT_LOG40 is unusable. Best to export the job, delete it, and re-import it.

The RT_LOG40 is a hashed type 30 file, the directory contained a DATA.30 and OVER.30 plus a hidden file, which are now gone and are not automatically created.

In a pinch you can do a complete copy from one of the other log directories into this one and then clear the log from the director.

Please, never just delete files from UNIX without being 100% certain of what you are doing.

Posted: Tue Sep 13, 2005 10:49 am
by Luk
Yes Arnd, you have completaly right - log hash is not recreated automaticaly.
Now in RT_Log40 directory are simple text files.

It was my quick solutuion! my second solution was that with exporting, deleting and reimporting job. Fortunately I have made backup before! I learned sth on my own mistake. Now I will do it as it should be. I will clear log in Administrator! :)

regards

Posted: Tue Sep 13, 2005 10:53 am
by Luk
by the way, ArnD, is there some documentation, or comand list (such as CLEAR.FILE), avalible to use from DS shell?