GOAL: To be able to run RGZPFM while production is running
COMMAND: RGZPFM FILE(TESTLIB/TESTFILE) RBDACCPTH(*NO) ALWCANCEL(*YES) LOCK(*SHRUPD)
RBDACCPTH was set to *NO to avoid the exclusive lock on the file
ISSUE: While running the RGZPFM, a couple records will briefly disappear and then come back
TEST CASE: I made an RPG program that ran the following steps
-Create a record with 6 million sequential records
-Delete the even numbered records
-Submit the RGZPFM command listed above
-Make "snapshots" of the file until it creates a file with less than 3 million records in it.
-When I analyzed the "snapshot" file, I found that roughly 20 records were missing out of the middle of the file.
THEORY: I assume this is happening because of the re-sequencing of the records?
QUESTION: Is there any way to avoid this behavior? My goal is to be able to run RGZPFM jobs while production is running, so missing records or exclusive locks are not an option for this.
Thank you
COMMAND: RGZPFM FILE(TESTLIB/TESTFILE) RBDACCPTH(*NO) ALWCANCEL(*YES) LOCK(*SHRUPD)
RBDACCPTH was set to *NO to avoid the exclusive lock on the file
ISSUE: While running the RGZPFM, a couple records will briefly disappear and then come back
TEST CASE: I made an RPG program that ran the following steps
-Create a record with 6 million sequential records
-Delete the even numbered records
-Submit the RGZPFM command listed above
-Make "snapshots" of the file until it creates a file with less than 3 million records in it.
-When I analyzed the "snapshot" file, I found that roughly 20 records were missing out of the middle of the file.
THEORY: I assume this is happening because of the re-sequencing of the records?
QUESTION: Is there any way to avoid this behavior? My goal is to be able to run RGZPFM jobs while production is running, so missing records or exclusive locks are not an option for this.
Thank you
Comment