PDA

View Full Version : Dbvisit Replicate - Trouble Shooting


Arjen Visser
08-24-2011, 12:09 AM
If Dbvisit Replication fails to replicate or the mine process is not able to connect to the apply process and vice-versa, there are a number of steps that can be taken to diagnose the issue.


Check that only one instance of each process is running for a given replication. Always kill all old processes.
Pay attention to all errors reported during the run of the setup script.
Check that the TNS configuration points to correct databases on all machines involved (mine, apply, console). Ensure there are valid TNS entries so that the apply database can connect to the mine database, and the mine database can connect to the apply database.
Ensure firewall ports are open to the default replicate ports (7890, 7891)
Ensure there is enough memory to start the mine and apply processes. The default memory allocation is 500MB. This is set by MEMORY_LIMIT_APPLY_MB and MEMORY_LIMIT_MINE_MB.
If the dbvrep process suddenly dies, it may mean that there is not enough free memory on the server and the OS will terminate the dbvrep process without warning. This can be verified by running "top" in Linux/Unix while the dbvrep apply or mine process is started.


Apply (or mine) cannot connect
|MINE IS running. Currently at plog 1071 and SCN 9832904 (06/10/2011 23:10:19).
Could not connect to APPLY process. Process not started or connection refused.


Check if the mine process is able to connect through the network to the apply process.
dbvrep> engine apply send engine status
-1: Connection refused
This indicates that there is a network issue. Check to ensure the firewall is not blocking the Dbvisit Replicate ports.
Run a health check command:
dbvrep> healthcheck
ERROR: Could not connect to APPLY, it is down or unreachable.
(Connection refused)
Connectivity test for APPLY: FAILED
ERROR: Could not connect to APPLY, it is down or unreachable.
(Connection refused)
Connectivity test for MINE: PASSED
2 processes checked.

Try ping to the unreachable process' server; try the ping from the mine machine, the machine console (if it is different from mine machine) and the apply machine itself. Check that the IP addresses reported by the ping are the same. (It definitely should not be 127.0.0.1 unless the replication is completely running on one machine.)

If the processes does not seem to see each other, try running “telnet machine port” on each of the machines to check connectivity to each other machine. The connection should not be refused. Use exactly same names as in your DDC file (use copy-paste if possible). This checks various DNS problems, typos in names/IPs, closed firewall ports etc.

Ensure there is enough free memory for the dbvrep process to start. f the dbvrep process suddenly dies, it may mean that there is not enough free memory on the server and the OS will terminate the dbvrep process without warning. This can be verified by running “top” in Linux/Unix while the dbvrep apply or mine process is started.