Traveler and Cluster (part2)

HCL Domino Domino Forum
1

I have two servers clustered.

In the case of a failure - users fail over to the other available server. Mail routing also fails over.

However Traveler does not regconise this fact.

The log file reports:

Notes Traveler: SEVERE CN=xxx/OU=xxx/OU=GB/O=xxx The Lotus Notes Traveler server cannot connect to your mail database mail/xxx.nsf on server CN=xxx/OU=xxx/O=xxx. Verify that your mail server mail database grants access to server CN=xx/OU=xx/O=xx and is operational.

The client cannot sync.

Because it cant see the users home server?. Can traveler on the MDA do the same as the client and fail over to the other cluster partner?

I can see that traveler has a traveler cluster cache. However for a user in here it references rep ids (that dont match the users mail file rep id) on one server, plus it says out of service. Is this similar to the cluster directory?

2

Subject: RE: Traveler and Cluster (part2)

You are going to need to open a PMR with the server activity logs for this problem to be debugged. Traveler should fail over in this case.

3

Subject: I have seen the same and have an open PMR for this

Also the cluster cache is cluttered with double entries with mixed case for the mail files.

4

Subject: RE: Double entries in Traveler Cluster Cache…

We have no clusters in operation, and have also witnessed that when users get a double entry, their device stops auto-sync-ing…

To solve the problem, since IBM seems unable to get us a fix until now we wrote an agent to delete the cluster cache entries every five minutes… since they’re always recreated anyhow… it doesn’t impeded the operation of traveler, and the lack of double entries keep our devices updating…

This really shouldn’t be necessary, but it’s the only workaround we could manage until now.