Start/notes.ini Parameters/REPL_PUSH_RETRIES

REPL_PUSH_RETRIES

🛠️
Parameter: REPL_PUSH_RETRIES
Short description: Number of times the replicator retries a failed push replication.

Profile

Parameter
REPL_PUSH_RETRIES
Category
Cluster
Component
Server
Available since
10.0
Supported versions
10.0, 11.0, 12.0, 14.0, 14.5, 14.5.1
GUI equivalent
notes.ini only (no GUI)
Possible values
Integer (default 3; typically 1–10)

Description

When a push replication run to another server fails (e.g. because of brief network problems between cluster nodes or a temporarily blocked target NSF), the replicator does not immediately start a new full run, but instead retries up to REPL_PUSH_RETRIES times within the same session. This absorbs flapping connections without restarting the entire replication cycle from scratch.
For stable LANs the default is sufficient; in WAN / VPN scenarios a slightly higher value pays off.

Example configuration

REPL_PUSH_RETRIES=5

Notes & pitfalls

  • Very high values can keep the replicator stuck on a broken connection longer – consider in combination with ReplicationTimeLimit.
  • Takes effect after a replicator restart (restart task replica).
  • Pairs with Server_Cluster_Probe_Timeout, ReplicationTimeLimit.
  • The cluster replicator has its own switches – this value primarily affects the classic Replicator.
  • For recurring push errors, first check the connection / ACL / quota before raising the retry count.