The Huawei Cloud Engine (Huawei_CE) vendor module no longer supports 'telnet' sessions to communicate with the device. This Huawei device family will hang up any incoming telnet session with the message it will not support this insecure protocol.

Additionally, the commit command is now executed only at the completion of a command job when meeting some restrictions. The commit is executed in case configuration has been changed, and the commit command is not included in the template or job.

After accessing the 'Management'-tab of a Node from the main Build menu, the 'Nodes' tab of the main build from would disappear. With several actions the tab could be made to re- appear, but this behaviour was a nuisance and was corrected.

When using the main Build general 'Search' to locate Sites, the result did not include Sites that had no nodes assigned. By including more result-types in the search could this issue be resolved.

The periodic polling for configuration changes of a node (nccm) could stop spontaneously. The NCCM node status would then indicate the polling was halted due to too many retries. And although the polling would proceed fine (when re-enabled), the true cause proved to be a Compliance rule.

When a 'command rule' is considered invalid by the node, the poll (which also executes these command rules) was considered failed. A few reties more and the node polling would be disabled.

This behaviour was changed such that only configuration retrieval failures will count as NCCM failures.

Attempting to install a recent NetYCE installation update would fail with an error message indicating a software library was missing. This would happen only on CentOS or RedHat version 6.x systems.

The reason was a dependency on a specific Perl library location that was changed with the update to a newer version. The dependency was resolved with the installation of the newer version, but was not available on the older systems.

The issue was resolved by testing for the dependency link before the NetYCE installation update process starts.

A similar library dependency exists for the MariaDB database. The installation was attempting to resolve this library even on systems where no database would be installed. This check was modified to be more specific.

Depending on the browser used, only three or four server buttons would be displayed in the NetYCE header. This issue is resolved by allowing room for up to six servers depending on server name length.

The service-type 'Add - Port - Type_by_name' is used to add an interface to a node based on the interface name it has on the device rather than the NetYCE abstract name. The function translates the interface name using the vendor-type.

As the Huawei CE devices now support 25, 40 and 100 Gbps interfaces using interface names like 25GE1/0/6, an issue was uncovered in this translation phase causing the function to fail. By correctly translating interface names starting with numbers, the issue was corrected.

A the NCCM periodically polls a node to backup and compare its configuration, it occasionally fails for one device or another. As these polling fails can have several causes (reachability, login credentials, permissions, transfer protocol, etc), correcting these failures need some feedback on which stage in the procedure the problem was encountered.

The cause of the failure could be found in the polling logs, but these are not easily found. To resolve this issue, the poll failure causes are now listed in the Comment field of the 'NCCM selection' tab data grid of the NCCM form.