Solarwinds - Setup

Im using solarwinds for system monitoring and have come to heavily rely on the alerting system which I have sending me text messages. Mainly just ups/downs, drive space warnings, things that need immediate attention.

I have it setup in a vm with its db located on our primary sql server. The problem is, if sql goes down or if everything goes down I obviously will not get these alerts.

If your using solarwinds this way, how do you have it setup? Anyone encountered this issue and what did you do about it?

I have two locations with the other location being more stable power wise. Thinking about exporting to OVF and migrating to my 2nd location DC then backup/restore the DB to a local sql express.

Thanks

Find more posts tagged with

Save $250 on 2025 certification boot camps from Infosec!

Book now with code EOY2025

Button

Comments

cruwl

We have ours running as VMs as well. Our DB server is just for SW nothing else running on it.

Not quite sure what youre asking, are you asking what do we do if we loose the entire data center? or if the sql server or Orion server go down?

tstrip007

I guess I was asking what the "best practice" setup would be. I may have answered my own question by needing to move it to its own DB. Do you have solarwinds db hosted on the same server?

Mrock4

SolarWinds recommends that the DB server is hosted off of the front end server- but if I recall correctly, they don't necessarily say you need a standalone device for the DB- just a separate box. In the several Orion installations I've worked with, the customers had a dedicated DB server which housed not only the Orion database, but several others.

DirtySouth

We are running SolarWinds NPM, SAM, NCM, NTA and Virtualization Manager all on a single VM with 6 processors and 8GB of RAM. The database is on a dedicated VM running on the same back EMC storage. As a fail-safe we do have a small instance of WhatsUp Gold running that monitors our critical servers and infrastructure. We are planning to add a second polling engine and split the monitored nodes between the two polling engines. That gives you a bit of redundancy and lightens the load on the primary polling engine.

Zartanasaurus

I run the DB on a remote SQL Cluster. If the DB goes down, I don't think that will change alerts since the DB is just for historical purposes. We have a modem attached to the Solarwinds server and most alerts come with a text message in case we lose the network.