Symptom
1. In HADR environment, Rep Agent is very slow and that caused ASE transaction log backlog (MB) keeps increasing in 'sap_status resource' output, as a result, primary ASE transaction log can not be truncated and HADR has to be disabled over and over again when log space is full.
2. Based on the collected information following KBA below
3171196 - How to collect information on performance with HADR 'Always On' - SRS
Multiple times of "select * from master..monProcessWaits" shows the event with the highest waiting time is WaitEventID=31 "waiting for buf write to complete before writing".
The only write operation performed by Rep Agent Scanner is the writing of the DBINFO to disk.
Here is a sample:
select * from master..monProcessWaits where SPID in (select ScannerSPID from master..monRepScanners where DBName='<SID>')
SPID InstanceID KPID ServerUserID OrigServerUserID WaitEventID Waits WaitTime
----------- ---------- ----------- ------------ ---------------- ----------- ----------- -----------
426 0 1764164700 0 0 29 81397 154090
426 0 1764164700 0 0 31 32804078 35526131
426 0 1764164700 0 0 36 104 185
426 0 1764164700 0 0 124 36228444 24718714
426 0 1764164700 0 0 150 12744 12896
426 0 1764164700 0 0 214 2004 33
426 0 1764164700 0 0 215 69155949 845197
426 0 1764164700 0 0 686 1 2016
426 0 1764164700 0 0 876 10 0
426 0 1764164700 0 0 908 29172 4516256
And this is confirmed by the stack traces collected for Rep Agent Scanner which shows the functions buf__writewait and dbinforelease.
Read more...
Environment
SAP Adaptive Server Enterprise with Always-on 16.0 SP04 PL08 and below
Keywords
KBA , BC-SYB-REP-SAP , Replication with SAP Suite / SAP BW , Problem
About this page
This is a preview of a SAP Knowledge Base Article. Click more to access the full version on SAP for Me (Login required).Search for additional results
Visit SAP Support Portal's SAP Notes and KBA Search.
SAP Knowledge Base Article - Preview