Project Corrupt

dspxguy · Post by **dspxguy** » Thu Oct 25, 2007 9:52 am

Hi we are getting this error message when trying to compile jobs or trying to create a copy.

Record xxxxxx in file DS_JOBS does not exist.

We did a DS.REINDEX ALL ( made sure no one was logged in to clients)

Then did

SELECT DS_JOBS WHERE NAME='JOBNAME';

0 records listed.

But when I do

SELECT DS_JOBOBJECTS WHERE NAME='JOBNAME';

I get this

Compiling "OBJTYPE".
FIELD (@ID, \, 1)
Compiling "OBJINDNO".
FIELD (@ID, \, 2)
Compiling "OBJNAME".
FIELD (@ID, \, 3)

Object type......J
Object record id....1112
Object record name...Root

1 record listed.

Why is it showing 0 records when doing select on ds_jobs?

Thanks

chulett · Post by **chulett** » Thu Oct 25, 2007 10:09 am

You have to use a real job name, not literally 'JOBNAME' in the select.

dspxguy · Post by **dspxguy** » Thu Oct 25, 2007 10:41 am

chulett wrote:You have to use a real job name, not literally 'JOBNAME' in the select. ...

yea in that case , I used the real job name :D

chulett · Post by **chulett** » Thu Oct 25, 2007 11:02 am

Then everyone was in fact not out of the project and the REINDEX failed. You should try again.

trokosz · Post by **trokosz** » Thu Oct 25, 2007 11:47 am

If there are any outstanding locks then the re-index does not work. Unlock then re-index.

dspxguy · Post by **dspxguy** » Thu Oct 25, 2007 11:59 am

trokosz wrote:If there are any outstanding locks then the re-index does not work. Unlock then re-index.

You guys might be right, I did check for the clients attaxched to the project, and there was none, but i did not check for any locks that were already held.

anyways, we will delete the whole project and create a new one and import the dsx from another dev box.

by the way, when we take a backup of a project which is by dsx and the other way we can do is by do a tar on the project folder from unix side, so if the project is corrupt and i take a tar of the corrupted project and then delete the corrupted project and make a new one with the same name and then un tar the project files in the new project...what will be the impact? how is tar and taking dsx different in DataStage when it comes to Universe internal table stuff?

please let me know your views

Thanks

dspxguy · Post by **dspxguy** » Thu Oct 25, 2007 12:02 pm

trokosz wrote:If there are any outstanding locks then the re-index does not work. Unlock then re-index.

You guys might be right, I did check for the clients attaxched to the project, and there was none, but i did not check for any locks that were already held.

anyways, we will delete the whole project and create a new one and import the dsx from another dev box.

by the way, when we take a backup of a project which is by dsx and the other way we can do is by do a tar on the project folder from unix side, so if the project is corrupt and i take a tar of the corrupted project and then delete the corrupted project and make a new one with the same name and then un tar the project files in the new project...what will be the impact? how is tar and taking dsx different in DataStage when it comes to Universe internal table stuff?

please let me know your views

Thanks

dspxguy · Post by **dspxguy** » Thu Oct 25, 2007 12:02 pm

trokosz wrote:If there are any outstanding locks then the re-index does not work. Unlock then re-index.

You guys might be right, I did check for the clients attaxched to the project, and there was none, but i did not check for any locks that were already held.

anyways, we will delete the whole project and create a new one and import the dsx from another dev box.

by the way, when we take a backup of a project which is by dsx and the other way we can do is by do a tar on the project folder from unix side, so if the project is corrupt and i take a tar of the corrupted project and then delete the corrupted project and make a new one with the same name and then un tar the project files in the new project...what will be the impact? how is tar and taking dsx different in DataStage when it comes to Universe internal table stuff?

please let me know your views

Thanks

trokosz · Post by **trokosz** » Thu Oct 25, 2007 12:11 pm

I don't believe your Project is corrupt and that you have to recreate...Just do a UNLOCK AL and then re-index and I'm sure your fine....

The tar backs up the filesystem which is the Project and all its files and you recover the entire Project.....The dsx backs up the Job designs, objects and so on within the Project itself and you can recover a single Job...Choose your poison

dspxguy · Post by **dspxguy** » Thu Oct 25, 2007 12:20 pm

trokosz wrote:I don't believe your Project is corrupt and that you have to recreate...Just do a UNLOCK AL and then re-index and I'm sure your fine....

The tar backs up the filesystem which is the Project and all its files and you recover the entire Project.....The dsx backs up the Job designs, objects and so on within the Project itself and you can recover a single Job...Choose your poison

Yeah that's right, but when you do a tar of the whole project and later on delete the project and recreate a new one with the same name then when you preserve the tar projects file in your new project, how will the jobs associate to the new project, becasue all jobs in the project are pointing to a project number something which is the old project.
I am not sure how does universe handle these things internally?

ArndW · Post by **ArndW** » Thu Oct 25, 2007 7:13 pm

If you restore the UNIX objects from a saved project directory to the same path then everything will work as most DataStage internal references are local and those few (but important) absolute paths used will also remain the same.