Change collector for Dynamic App in device

Question

We have this kind of issueDevice is monitored with SNMP and also by Juniper Mist system. So actually there is first two separate devices in SL1. Because that device is a switch stack of for physical switches which also need to be discovered. The stacked devices discovery is pretty simple for us, we create component device for each physical device and get serials and firmware versions etc (CMDB requirements).But when we merge then this goes a bit messy. SNMP monitoring is done from a CUG in customer network, Mist connection is done from separate CUG which has access to Internet. All basic stuff is fine, but the component devices (those physical switches) will have interesting situation:the component device it self is mapped correctly to customer CUG, which is needed to get snmp access to that networkDA will be aligned to backend CUG, the one which is used for connecting to Mist, and which of course does not have access to customer CUG/collector where the cache is stored.So, is there any way to force that DA to use same CUG/Collector that is seen in device? Logically in UI is shown the default, but DA alignment will choose the other one.

tonyandres · Answer

Hello Teppo,
In the System &gt; Manage &gt; Dynamic Applications page, if you click on the wrench icon for the dynamic app you want modified, you can change the Collector Affinity in the Properties tab. You have the following 3 choices:

Default. If the Dynamic Application is automatically aligned to a component device during discovery, then the Data Collector assigned to the root device will collect data for the Dynamic Application. Otherwise, the Data Collector assigned to the device running the Dynamic Application will collect data for the Dynamic Application.
Root Device Collector. The Data Collector assigned to the root device will collect data for the Dynamic Application. This guarantees that a group of collection jobs that cannot be split up run on a single Data Collector. You might select this option if this Dynamic Application has a cache dependency with one or more other Dynamic Applications, if you are unable to collect data for devices and Dynamic Applications within the same Device Component Map on multiple Data Collectors in a collector group, or if the Dynamic Application will consume cache produced by a Dynamic Application aligned to a non-root device (for instance, a cluster device).
Assigned Collector. The Data Collector assigned to the device running the Dynamic Application will collect data for the Dynamic Application. This allows Dynamic Applications that are automatically aligned to component devices during discovery to run on multiple Data Collectors. You might select this option if this Dynamic Application has no cache dependencies with any other Dynamic Applications, if you want the Dynamic Application to be able to make parallel data requests across multiple Data Collectors in a collector group, or if the Dynamic Application can be aligned using mechanisms other than auto-alignment during discovery (for instance, manual alignment or alignment via Device Class Templates or Run Book Actions).

For the example you provided above, you may want to set the affinity to Assigned Collector so that the dynamic app will always be aligned to the same collector as the device itself. Setting it on the dynamic app level will always select this affinity when SL1 aligns this specific app.
Antonio Andres
Principal Technical Support Engineer | ScienceLogic

teppotahkapaa · Answer

Hi Antonio,well, my issue has been with devices that are merged from two devices which are in separate CUGs. And when you align DA to device then SL1 makes the decision to which CUG it puts it. And whatever I try it always goes to wrong CUG. So when merged device is:device1 (snmp monitored, in CUG1)device2 (API monitored, in CUG2)then always the snmp DA I try to align there goes to CUG2, obviously SL1 thinks that the device with bigger DID is the default, and this case also root, and of course the one that does not have snmp access to that network.Teppo Tähkäpää

tonyandres · Answer

Hello Teppo,
When you merge a device, SL1 has to pick one device as a "master" device, which is usually the component device as it still needs to be part of the component tree. You cannot have a device aligned to 2 collectors so the 2 merged devices started out on different collectors, the new merged device will be placed in 1 of the collectors, normally the one where the component device was aligned. It is up to you to ensure that both collectors can reach the device via whichever protocols are necessary to monitor the device. You cannot force a dynamic app to be aligned to a collector in a different CUG, as this goes against the architecture of SL1. Only collectors in the same CUG will have visibility on the devices in that same CUG, not a different CUG. This scenario you are describing is not supported in SL1.
Antonio Andres
Principal Technical Support Engineer | ScienceLogic

teppotahkapaa · Answer

So, making that clause more human-understandable. Merging of devices from two different CUGs is only&nbsp; supported if all same access to monitored devices/systems from all collectors from both CUGs is available ?

tonyandres · Answer

Hello Teppo,
That would be accurate.
Antonio Andres
Principal Technical Support Engineer | ScienceLogic

Forum Discussion

Change collector for Dynamic App in device

5 Replies

Recent Discussions

ServiceNow CMDB SyncPack v3.6.3 Release

Dashboard filter using output

Power flow Sync Pack failures

PowerFlow debugging

Dynamic device group automation with GQL