Page 1 of 1

Error in Compling Server jobs

Posted: Mon Apr 14, 2008 1:23 am
by tkbharani
When I am compiling any of my server job it throws an error " Error Calling RTLog<numbers>". Some times it says "check data stage is set up correctly"

Why is this occurring. Is rebuilding job index is the solution. I have searched forum but did not get solution.

Posted: Mon Apr 14, 2008 1:43 am
by ray.wurlod
Please post the exact message.

Check also that your project directory's file system is not full, and that you have write permission to it.

Re-indexing is unlikely to help, as the log tables are not indexed.

Posted: Mon Apr 14, 2008 1:53 am
by tkbharani
ray.wurlod wrote:Please post the exact message.

Check also that your project directory's file system is not full, and that you have write permission to it.

Re-indexing is unlikely to help, as the log tables are ...
Yes, I think my file system is full its almost 96% used. So what files I to delete from my project directory to bring it down.

Posted: Mon Apr 14, 2008 2:14 am
by ray.wurlod
Anything in &PH& directory would be a good start. Clear the &PH& directory in all projects. Any file in &SAVEDLISTS& directory that has a name beginning with "&&S." is also not required.

Are there any temporary files (or temporary hashed files) that have been created in the project directory? If you are planning delete hashed files, beware that there's a right way and a wrong way - and using the operating system to delete hashed files from a project is the wrong way.

Allocate some time to checking your job logs. These really do need to be kept as small as possible, commensurate with your retention policy. If a job log is huge, look at it in Director, observe that there are millions of warnings, and purge old entries. If you want to reclaim some of the job log disk space, you will need to RESIZE job log tables.

A useful set of commands for gathering log file sizes are these:

Code: Select all

SELECT VOC WITH TYPE = "F" AND @ID LIKE "RT_LOG..."
ACCOUNT.FILE.STATS ALL LOCAL
These will, together, gather sizing information about all your log tables into a table called STAT.FILE. You can query this in various ways to identify the "hogs", for example

Code: Select all

SELECT * FROM STAT.FILE;
or even the pre-stored query

Code: Select all

LIST.FILE.STATS LOCAL