"Merging" may not be quite the right word but that is the desired end result.

Scenario: many Solaris 10 servers, each with various local users. We want to set up LDAP for all for all of them. LDAP server is set up, procedure for getting other servers to use it for user authentication is documented and tested. The question is how to handle users that are in LDAP who also exist as a local user on a given machine.

It appears that the usernames on both sides follow a convention and therefore match but obviously the userids will not match. Local user joe has userid 1234, LDAP user joe has userid 56789.

The way I see it we'll have to:
1. move local user joe's home directory to the path that LDAP user joe will want
2. change local user joe's userid to that of LDAP user joe
3. change joe's files' owner to his new userid
4. remove local user joe
5. finally configure LDAP

Is this a rational procedure? Is there a more effective method? I'm not looking forward to this as there are many servers and each of them have a different set of local users, each with different userids which will have to be handled manually and individually therefore not even scriptable much.

TIA

Your procedure is assuming user joe isn't logged in during the migration, and has no processes running under his account. If that is the case, it seems correct although you are missing to handle possible group mismatches.

If the user account need to stay active and usable without disruption during the migration, you might use ACLs to grant the very same rights on files belonging to joe to both 1234 and 56789 user ids.