3437826 - log buffer takes more than 30000000 us to process and write, this could be caused by write-IO peak

In a System Replication landscape, you may notice frequent log shipping timeouts or even log volume full situations occur for secondary site.

In the indexserver trace on secondary, you see the entry

w PersistenceManag DisasterRecoveryProtocol.cpp(04582) : log buffer takes more than 30000000 us to process and write, this could be caused by write-IO peak.

w PersistenceManag DisasterRecoveryProtocol.cpp(04583) : This trace is intended to clarify frequent "logshipping timeout" situation. If this is the case, then it's probably by write-IO, not by network.

SAP HANA with system replication enabled

io , KBA , HAN-DB-HA , SAP HANA High Availability (System Replication, DR, etc.) , Problem