Bug in version 223 related to netscan execution
Reporting a bug we have identified that was introduced in portal version 223. May affect any portal that is using scheduled netscans to discover monitored devices.
Starting in portal v223, if a netscan runs and discovers a device which is already in monitoring, it will re-assign that device to the collector which is running the netscan.
In our case we have many scans which run each night against different subnets. We have one collector running the scans, and the collector is part of an auto-balanced collector group so the devices get distributed once discovered.
Starting with this bug, whenever the nightly scans run it will reassign many devices to the single collector, which then gets overwhelmed and monitored devices generate false Host Status alerts. Auto-balancing only helps to a certain extent because the balancing is based on instance count, not resource count
We have a workaround in place and are working with LM support to get the underlying bug resolved, but wanted to make the community aware.