Recovery
In the event a server is down or failed there are different recovery actions needed depending on the type of failure:
• Server down:
If one of the servers involved in replication is down for a relatively short period of time, simply restart the server. The missing changes will automatically be replicated and applied and then all servers will be in sync. No other manual steps are needed.
• Server down for a long period of time:
If one of the servers involved in replication is down for a day or longer, data may need to be re-synchronized. This request will cause the server’s data to be rebuilt from data on the specified server.This can be done using a command similar to the following. Modify parameters as appropriate:
dsreplication initialize --hostSource <server-that’s-running-well> --portSource 4444
--hostDestination <server-that-was-down> --portDestination 4444 --adminUID admin
--adminPassword admin --baseDN o=ptc --trustAll --no-prompt
• Server fails totally:
If one of the servers involved in replication fails completely due to a hard drive failure or other such failure, recovery should still be relatively easy. Correct the underlying failure and simply re-install WindchillDS on the failed system. Then, re-run the dsreplication enable command as for the initial setup. Then run a dsreplication initialize command similar to that in the previous paragraph. All servers should again be in sync.
• Servers no longer in synch:
If one of the servers involved in replication is no longer is sync with the other servers, run a dsreplication initialize command similar to that in a previous paragraph.