• Article Type: General
• Product: Aleph
• Product Version: 20, 21, 22, 23

Description:
How many processes should I specify in running p_manage_01 and p_manage_02?

Resolution:
The general issue is discussed in Section 4 ("Processes and Cycle Size") of the "How To Run Index Jobs" document (in Ex Libris Documentation Center > Aleph > Support > How To from Support by subject > Indexing_filing_and_expand_procedures ).

I'm adding the following to Section 4:

In regard to the p_manage_01 or p_manage_02 jobs, which have single-threaded sections:

In the case where the multi-threaded p_manage_01_a extraction processes have "run ahead" of the single-threaded p_manage_01_c, p_manage_01_d, and p_manage_01_d1 processes, you will find, in the $data_scratch directory, after the job has completed, that the last p_manage_01_a_n.log has a timestamp 15, 30, or 60 minutes earlier than the final p_manage_01_c.log timestamp. In this case you want to reduce the number of p_manage_01 processes in any subsequent runs of the job.

If, on the other hand, you find that just two or five minutes separate the timestamp of the last p_manage_01_a_n.log and the final p_manage_01_d1 timestamp, that's an indication that the number of p_manage_01 processes can and should be increased. Another indication: if you find that the plus signs in column 1 of the cycles file are not running ahead of the other columns -- that is, the other processing is keeping up with the extraction --, then you may want to increase the number of processes in subsequent runs in order to get maximum performance.

• Article last edited: 10/8/2013