Salesforce

LOCKED ai extents are not archived by AIMGT

« Go Back

Information

 
TitleLOCKED ai extents are not archived by AIMGT
URL NameP124750
Article Number000145410
EnvironmentProduct: OpenEdge
Version: 10.1A, 10.1B
OS: All supported platforms
Question/Problem Description
AIi extents with LOCKED status are not archived
When ai extents are LOCKED, the aiarchiver doesn't archive these
aiarchiver continues to switch ai extents as per the setinterval n or as ai extents fill
source database eventually runs out of available ai extents to switch to 3775 3776
Steps to Reproduce
Clarifying Information
Error MessageAIMGT   6: Switched to ai extent . (3777)
AIMGT   6: (3775)  Can't switch to after-image extent \repl\source\source.a1 it is full.
AIMGT   6: (3776)  Backup ai extent and mark it as empty.
Defect NumberOE00154775
Enhancement Number
Cause
LOCKED ai extents are not considered FULL in OpenEdge 10.1B and 10.1A.
Resolution
Upgrade to OpenEdge 10.1C where the aiarchiver will archive LOCKED extents.

The AI Management Daemon will archive LOCKED extents, however they are marked as ARCHIVED and not available for re-use.
      
AIMGT   6: (3777)  Switched to ai extent dbname.a1. 
AIMGT   6: (3778)  This is after-image file number 25264 since the last AIMAGE BEGIN 
AIMGT   6: (13199) After-image extent dbname.a7 has been copied to <-aiarcdir>. 
AIMGT   6: (13154) Marked after-image extent dbname.a7 ARCHIVED.

The RFUTIL -C AIMAGE LIST report will still show these " ARCHIVED" ai extents as "LOCKED".

When replication resumes and they have been fully processed against the target database, the AI Management Daemon will mark the extent as EMPTY for re-use. These extents will not be backed-up as they were backed up previously at the time they were first switched and marked as "ARCHIVED".  
      
RPLS    5: (11805) Unlocking after-image file 25263 and locking ALL FULL after-image files beginning with file 25264. 
AIMGT   6: (3789)  Marked after-image extent dbname.a7 EMPTY.

Should replication not resume before all ai files in the sequence chain have been exhausted, the database with either shutdown or stall depending on if the -aistall parameter was set when the database was initially started.  To recover review the Options described in Article  Replication: source database does not restart due to error 3773 5350 on ai file  

Until Replication resumes in any OpenEdge version:

Ensure there are available AI extents to record transaction activity.

1.  Increase the aiarchiver interval or set it to "on demand" with the value 0, to preserve ai extents until replication resumes:
$   rfutil source -C aiarchiver setinterval 0

or 

Stop the aiarchiver deamon: 
$   rfutil source -C aiarchiver end

Once repilcation resumes, the aiarchiver deamon can be restared with the following command line:
$  _mprshut source -C aimgt -aiarcdir xxxx -aiarcinterval n [-aiarcdircreate]

2.  It may be necessary to add more ai extents to record the database activity.

3.  If there is concern that the primary server will fail, these LOCKED ai extents can be manually copied to an archive directory on another server with OS utilities in the meantime.
Workaround
In versions prior to OpenEdge 10.1C, when replication resumes, and these LOCKED extents are processed to the target database and marked FULL, the aiarchiver daemon will archive these in sequence then mark them as EMPTY.

RPLS    5: Unlocking after-image file <num> and locking ALL FULL after-image files beginning with file <num>. (11805)
AIMGT   6: After-image extent <name> has been copied to <-aiarcdir>. (13199)
AIMGT   6: Marked after-image extent EMPTY. (3789)
 
Notes
Keyword Phrase
Last Modified Date11/20/2020 7:37 AM

Powered by