- Article Type: General
- Product: Aleph
- Product Version: 18.01
The last time we ran manage_07 in abc01 was 01/15/08. We want to run it again 12/22/08, since it's supposed to be run "peridocially." We want to rebuild the z00Rs at the same time we rebuild the z13s, as one of our libraries wants to use the z00Rs more for statistics. Last time manage_07 took 2 hours to run. How much longer should we anticipate it running if it's also going to do the z00Rs this time?
There is no need to periodically run p_manage_07. It only needs to be run when tab22 changes -- or when an upgrade requires it.
But it seems that p_manage_07 has not been run for the abc01 z00r -- since I find there are only 95,000 z00r's (versus 5.3 million z00's). (There should be about 15 times as many z00r records as z00 since each z00 field becomes a separate z00r record. In contrast, there is only one z13 record for each bib record.)
The older timings I have of p_manage_07 are from when it did just the z13. The number of z00r records being written is about 15 times as great as the number of z13's. I believe that doing both the z13 and the z00r will make the run time 3 - 5 times as long as a run for only the z13, but I'm not sure.
Since (unlike p_manage_01 and p_manage_02) p_manage_07 has no single-threaded steps, it benefits from having a large number of processes. If the job is being run during off-hours, you could try maybe 16 processes.
See also KB 8192-5719.
- Article last edited: 10/8/2013