p_manage_01_e: How To Restart Load of z98
- Article Type: General
- Product: Aleph
- Product Version: 16.02
Description:
If p_manage_01_e fails in the p_manage_01_e2 (load z98 step) after all the $data_files/z98.seq files have been created, is there some way you can restart just this load step?
Resolution:
[This information is also in section 9 ("Troubleshooting") of the "How To Run Index Jobs" document. Note: You can do this ONLY if the z98.seq (or z02.seq) is complete. If it is not complete, more elaborate steps -- as described in the document's preceding "Restart of p_manage_01_e" section -- would be required.]
Re-Run the z98 Load:
1. UTIL-A-17-1 for z98, then UTIL-A-17-4 (to delete the z98 index, since z98 is loaded with no index).
2. Save a copy of the $data_scratch/p_manage_01_e.cycles file in case of error:
>> cp -p p_manage_01_e.cycles p_manage_01_e.cycles.save
3. Edit the original $data_scratch/p_manage_01_e.cycles file, changing column 2 from "+" to "-":
:%s/ + +/ + -/g
4. Either:
a. place a util_e_stop file in the xxx01 $data_scratch (as described in section 7 of "How To Run Index Jobs")
OR
b. do UTIL-C-5 to lock the BIB library.
5. Run this command to reprocess the z98.seq files:
a. v14.2
>> csh -f $aleph_proc/p_manage_01_e2 XXX01,1,000000000,999999999,,9 >& $data_scratch/p_manage_01_e2_recover.log &
OR
b. v15.3 (and, I believe, 15.2:)
>> csh -f $aleph_proc/p_manage_01_e2 XXX01,1,000000000,999999999,,9 p_manage_01 >& $data_scratch/p_manage_01_e2_recover.log &
6. Run UTIL-A-17-2 for the z98_(to build the z98_id Oracle index).;
This is necessary in the case of a remote database / "two-task" setup. But in the normal, non-remote case it may be possible to load these sequential files using p_file_04. For the z98 you would do:
>> mv z98.seq.1 z98.seqaa
>> mv z98.seq.2 z98.seqab
etc.
And then, to load them:
>> csh -f $aleph_proc/p_file_04 xxx01,z98,replace
... But we know *for sure* that the preceding works.
(Keywords: p-manage-01 manage_01 manage_01_e)
- Article last edited: 10/8/2013