Search the Community
Showing results for tags 'constant change'.
Found 1 result
My problem is pretty simple. I have 10,000 USB hard drives sitting close to one of my servers. Every 5 or so minutes we have a robotic hand attached to a robotic arm, which is attached by USB if that's important, that unplugs a random number of the USB drives that are connected to the system, there are only 16 USB ports after all, and it then plugs new USB drives into each of the now vacant USB ports. So what ends up happening here is that upon an autodiscover new instances are created for each of these new drives that have been attached to the system. These instances inherit their alert tuning from the datasource. In this case that is snmpHRDisk-. At the time of this writing snmpHRDisk- is a datasource associated with 158 hosts. If I change the settings to satisfy this host's alert tuning, applies to or filters these apply to all hosts. Changing the alert tuning to satisfy this host's needs virtually disables alerting on all hosts. Changing the filters might cause us to miss a similar naming scheme on another host. So far the best option at present seems to be duplicating the datasource, which we did, then changing the applies to in the main datasource. When I went to change the applies to I've so far managed to associate the datasource to all 1000+ of our hosts, no hosts, only the two hosts, etc... I think the best solution to my problem is the ability to simply disable a datasource at the device level and/or allow alert tuning to be set on the parent of the instances in the device so that new instances will inherit their settings from the device and not global and/or allow new instances to be added to a predefined instance group with alert tuning. There may be different ways to skin a cat but disallowing me from skinning the cat on moral grounds is just immoral. But this one wasn't intelligent. He blew all his money on lottery tickets!