The install of the way you show me ¿is it for loss of service?,
As you show me would be:
1. - Passive Node: install SP2 CU1
2. - Balancing cluster with instances. -- What do you mean by balancing as SQL server does not support load balancing?
3. - Check that the passive node instances are in: 10.50.4260 -- You will need the node to be active to check the version, when passive you can only check in the installation log created by the CU update.
4. - Active Node: Installing SP2 CU1
5. - Balancing the active node. -- Same question as above.
6. - Check that the active node instances are in: 10.50.4260
Is it right?
I'm sorry but I am having real difficulty understanding your descriptions. If I am interpreting your question correctly you will only experience downtime of the SQL server during the initial failover to a patched instance, this time will incude the taking offline of the SQL Server and its resources from the active node and bringing them online on the passive node. In addtion once the resources are back online the update process will automatically upgrade the system databases.
There is no exact figure on how long this will take but on our production SQL servers which have approx 15 databases with a total storage of 4TB this took approximately 6 minutes, During this time the server and databases are unaccessible.
When failing back from the now patched active node to a patched passive node the time should be less as the databases have already been upgraded during the first failover.
MCITP SQL 2005, MCSA SQL 2012