Consultor Eletrônico



Kbase P144784: Replication agent dies with SYSTEM ERROR: lkrels record not locked. (435)
Autor   Progress Software Corporation - Progress
Acesso   Público
Publicação   18/01/2010
Status: Verified

SYMPTOM(s):

Replication agent dies with SYSTEM ERROR: lkrels record <recid> not locked. (435)

Replication agent will crash when processing recid over 4 billion range

Replication agent crashes when processing recid over 4 billion (2^32) range

recid reported by the 435 message is the id of the record recently created on target database /minus/ 4 billion.

After the 435 errors appear, the target crashes with error 6495 5028

Out of free shared memory. Use -Mxs to increase (6495)

SYSTEM ERROR: Releasing regular latch. latchId: 14 (5028)

latch 14 is "Lock Table Free List" in error 5028

Synchronisation fails after target is restarted 1028 8999

Replicaiton synchronization point fails with following errrors:
SYSTEM ERROR: Rollforward Apply AI note, BKUPDCTR=2, note updctr=1. (1028)
area <areanum>, dbkey 4294967296, record type <recordtype> (8999)
Function dsaAIApplyNote failed in rpNOT_ProcessNoteBlock with error -30000. (10428)
Block processing ended in error at block <block> in area <areanum>. (10495)
Unexpected error -30000 returned to function rpAGT_AgentLoop. (10504)

after target database is recreated from a source backup, replication agent continues to throw error 435

FACT(s) (Environment):

OpenEdge 10.1x
OpenEdge 10.2A
All Supported Operating Systems

CAUSE:

Bug# OE00181622

CAUSE:

A product limit: OE replication crashes when the High-Water Mark (HWM) in an area reaches 4 billion (2^32). Furthermore, each error 435 causes 64 bytes of memory leaks. As result the target is periodically (each a few hours) crashed with error 6495

FIX:

Upgrade to 10.1C04 or 10.2A01 or higher OpenEdge release where this issue has been addressed.

Workaround:

Re-organise the database schema by moving objects out of the storage area affected, into new storage area(s) so that the 4 billion rowid's is not reached. Then remove this storage area and re-baseline replication (ie: re-create the target database)