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

Description:
Can't log on to Web in early morning (after clear_vir01 has been run) www_server log shows: Oracle error: io_z63_write ORA-00001: unique constraint (VIR01.Z63_ID) violated

Resolution:
[Note: See KB 5297 for a discussion of this issue, in version 17-up.]

This error was logged at 06:00: the exact time that clear_vir01 was running. So someone was trying to log into the Web as clear_vir01 was running. This is the problem described in PRB 4058. The difference is that the 16.02 clear_vir01 proc has been suppressing the "details" (including any errors!). 16.02 rep_change 547 changes clear_vir01 to create the index for each table immediately after creating the table itself. (This true not just of the z63 but each vir01 table.)

This narrows the window for a possible problem from about 20 seconds to about 1 second. It also makes the detailed log be written (except when run through util x/8 -- in that case: no detailed log).

If the problem is occurring in a pre-16.02 version, you can change the $aleph_proc/clear_vir01: After the " csh -f $aleph_proc/util_a_17_b index all " line add these lines: echo " " echo "Recreating z63 table and index" csh -f $aleph_proc/util_a_17_b table z63 csh -f $aleph_proc/util_a_17_b index z63. You can change the clear_vir01 to run at a time when fewer people would be likely to log in. In the unlikely event that even with the preceding measures you still see the problem, you can do > dlib vir01 > util a/17/1 <for the z63> to correct it. Another approach would be to have a script stop the www_server while clear_vir01 is running (and restart it when done).

See KB 5297 for a discussion of this issue, in version 17-up.

[Note: I have intentionally *un*-published this record because it is for version 16.02 and it has been superceded by other records. Jerry ]