1.13. Release Notes 1.5.2¶
1.13.1. Improvements and New Features¶
Support for discovery and monitoring of Brocade VDX devices has been added. This includes network topology discovery in VCS Fabric configuration.
Multiple improvements in the discovery of network topology using information from LAG and LLDP protocols. NetSpyGlass will not try to add links between MLAG port channel interfaces to maps.
Implemented discovery of COS queues on interfaces of Juniper QFX devices.
Added monitoring of jnxCosQstatTotalDropPkts on interfaces of Juniper QFX devices
Cluster configuration validation:
- “Data push loops” are not allowed: cluster configuration where server ‘server1’ is configured to push data to ‘server2’, while ‘server2’ is configured to push back to ‘server1’, can not be loaded. Documentation has been updated to reflect this limitation.
- cluster configuration where a server or monitor is configured to push data to a cluster member that is not defined can not be loaded.
- parameter ‘allocation’ is only allowed for cluster members with role ‘monitor’
changes in the contents of the file cluster.conf do not trigger full reconfiguration, they only cause device rebalancing between monitors and force monitors to reload devices and variables.
multiple changes to streamline startup sequence and to avoid unnecessary device rebalancing on startup
common parts of the configuration are now partially stored in zookeeper (cluster configuration and common monitor config, such as polling interval, snmp timeouts, retries, batch size etc)
1.13.2. Bug fixes¶
- secondary server should not try to make data push to the primary for devices for which it has no variables
- restored ability to color links that terminate on network devices we do not monitor, e.g. external peering links