Killing p_manage_05 in the middle of the process?
- Article Type: General
- Product: Aleph
- Product Version: 18.01
Description:
I submitted a p-manage-05 job last night on our test server and it is still running after 24 hours. So far it has only done 1/3! I am afraid of affecting others' work because it has locked MGU01 and am thinking about killing the process, but I am not sure if I can do it in the middle of the process.
(I used the default option: Delete entire index and rebuild for range of documents). What should I do now?
Resolution:
You submitted p_manage_05 with only one process:
setenv p_active_library "ABC01"
setenv p_delete_flag "1"
setenv p_start_doc_number_x "000000000"
setenv p_end_doc_number_x "999999999"
setenv p_no_process_x "01" <<-----------------------
Though having too many processes can be a problem for p_manage_01 and p_manage_02, p_manage_05 and other indexing jobs benefit from having as many processes as possible.
The "uname -X" command shows that you have 11 processors. The typical rule for indexing processes is (# of processors) * 2. That would be 22 -- but the maximum is 20. I suggest that in the future you run p_manage_05 with 16 processes.
The job is nearing the end (just five cycles out of 54). You need to let the job finish. If you do not, you will have no Direct index at all.
- Article last edited: 10/8/2013