Solve Replication Processing Problems

Diagnose, monitor, and solve possible problems that can occur while Enterprise Replication is running.

You should understand the typical behavior of your Enterprise Replication system. There are many factors that contribute to the performance and other behaviors, including: hardware configuration, network load and speed, type of replication, and number of replicated transactions.

Use the cdr view command or the SMI tables to understand the typical behavior of your system, establish benchmarks, and track trends. Deviations from typical behavior do not necessarily indicate a problem. For example, transactions might take longer to replicate during peak usage times or during end-of-month processing.

The following table describes some replication processing problems that might occur.

Table 1. Potential Replication Problems and Solutions

Problem	How to diagnose	How to solve
Enterprise Replication is not running	Run the cdr view state command Query the syscdr_state SMI table Examine event alarms captured by the alarm program	Start replication with the cdr start command.
One or more Enterprise Replication servers are not running or connected to the network	Run the cdr view servers command Run the cdr view nif command Query the syscdr_nif SMI table Examine event alarms captured by the alarm program	Start the database server or fix the connection problem.
Replicated transactions failed	Determine if there are ATS or RIS files: Look at the ATS and RIS directories on the local server for the existence of ATS or RIS files Run the cdr view atsdir risdir command to see the number of ATS and RIS files for each server Query the syscdr_atsdir or syscdr_risdir SMI table for a specific server Examine event alarms captured by the alarm program	Run one of the following commands: cdr repair cdr check replicate --repair cdr check replicateset --repair See cdr repair, cdr check replicate, and cdr check replicateset.
Transactions are spooling to disk	Determine how much spool memory is being used: Run the cdr view profile command to see the status of all queues on all servers Run the cdr view sendq command to see the status of the send queue on all servers Run the cdr view rcv command to see the status of the receive queue on all servers	See Increasing the Sizes or Numbers of Storage Spaces.
Potential log wrap situation	Determine how many log pages must be used before Enterprise Replication reacts a potential log wrap situation: Run the cdr view ddr command to see the number of unused log pages for all servers Query the syscdr_ddr SMI table to see the number of unused log pages for a specific server	See Handle potential log wrapping.

If you do need to call HCL Software Support, find the version of the database server that is running Enterprise Replication with the cdr -V command.