The database YCE has had its tables changed from the MyISAM database schema to the InnoDB database schema. The other databases remain using the MyISAM schema. This should not affect performance or database replication.

Service types now are rolled back entirely when a failure occurs in one of its steps. In the past, you had to clean up the different records that were created or modified in the previous steps manually. With this new change this happens automatically. This is the reason why the YCE database had to be changed to use the schema of InnoDB, as MyISAM does not support this feature.

The one exception to this is changing the hostname of a node in a service type. Any records changed in the NCCM database will not be rolled back. Unfortunately it was not possible to also change the NCCM database back to using the InnoDB schema. If a service type that changes the hostname of a node fails in one of its steps, any record in the NCCM database has to be changed back manually.

Two new features for the service types have been added. ADD SUBNET now has a FIRSTFREE_SUBNET_OFFSET option. This is a derivative of the FIRSTFREE_FROM_SUBNET service type, but it allows you to concatenate a number after the subnet name, which will take the first free subnet starting from that offset in the provided subnet. So for example, the value of '<mgmt> 12' will take the subnet in the <mgmt>-alias, and start looking for the first free address after skipping the first 12 addresses.

In a similar vein, the service type LOCATE ADDRESS FIRSTFREE_ADDR has been modified so that its value also accepts an offset that allows you to search for the first free address in the subnet, starting from that offset.

Improved regular expression so the software version gets extracted for the following JunOS node models as well: ex2200-c-12t-2g ex4200-24f ex4550-32f

Added code to the Cisco IOS vendor module which makes it possible when you run the “install remove inactive” command to take some time before the confirmation prompt will be answered.

When you delete a node directly with a database query, the nccmd daemon would detect this and clean up any records in backups selection. It would not clean up compliance records. This is now fixed.

The skulker now supports nested directories in the OS repo

Now extracting the actual node model for Cisco WLC

Fixed the sftp put subroutine for the Checkpoint vendor module

There was an error in a scenario generated by the RN3 failover tool. This is now fixed

Due to timing, searching for reports sometimes did not show all results. This is now fixed, but may have made the reports grid respond a bit slower.

The md5 was not properly calculated for a new OS file in the OS image form. This is fixed. Also, at times the main grid was not properly updated after changes. This also fixed.