Skip to main content
Ex Libris Knowledge Center

Restarting p_manage_01

  • Article Type: General
  • Product: Aleph
  • Product Version: 16.02

Description:
Apparently a "network glitch" occurred around 11 AM this morning effectively suspending the p_manage_01 job that had been running since Monday at 3:30 AM. It would appear that the job had completed through cycle 63 (record 003150000).

Is it possible for me to start the job again with the FROM DOCUMENT NUMBER set to 003150001 to finish it out or does it have to be started from the beginning again?

Resolution:
Following a similar procedure for p_manage_02, described in KB 8192-2231, we restarted p_manage_01 as follows:

[Note: The following should be done only in consultation with Ex Libris Support.]

1. Save the current p_manage_01 as p_manage_01.normal:

cd $aleph_proc
cp -p p_manage_01 p_manage_01.normal

2. Insert two lines into p_manage_01:

if (`gs tab00_correct` != 00) then
echo "Error(s) in tab00, EXIT"
goto ex_p_manage_01
endif

goto restart <----INSERT
$aleph_proc/aleph_plus @$aleph_proc/correct_last_doc_number.sql $p_active_library
...
...
cd $data_scratch
rm manage_01*
rm p_manage_01_*.log
restart: <---INSERT
cd $aleph_proc

3. create a "p_manage_01.1strun" directory in the abc01 $data_scratch directory and move all of the p_manage_01* files to it. (Note: This is *just* the p_manage_01* log files; *not* the manage_01* files.)

4. copy the p_manage_01_a.cycles file from that directory back to $data_scratch and then edit the actual version (in $data_scratch) as follows:

Change all the "*" to "-".
Change all the "!" to "-".
Change all the "?" to "-".

[In this case, there were no cases of "!" in columns 2-4. If there are, there *may* be a potential for the restart creating certain duplicate z97/z95's. Though this may not really be a problem.... ("!" in Column 1 can be ignored; that is never a problem -- though, as noted above, it should be changed to "-".)Columns 2, 3, and 4 of the p_manage_01_a.cycles are "single-threaded", that is, the job will only be working on one column for one cycle. Column 2 is the writing of the Z97's. Column 4 is the writing of the Z95's.]

5. (Since the previous p_manage_01 failure will have left the library locked) use util c/6 to unlock the library.

6. Submit p_manage_01 with *exactly* the same parameters as were used in the original submission. (If you submit it from the command line, you can add a suffix of "restart" to the log file name.)

7. After the p_manage_01 restart has been submitted, restore the regular p_manage_01 proc:

ap
mv p_manage_01 p_manage_01.restart
mv p_manage_01.normal p_manage_01

8. This information will be added to the "Restart" section of Chapter 9 of the"How To Run Index Jobs" document.


  • Article last edited: 10/8/2013
  • Was this article helpful?