I’ll jump first, it builds self esteem.
I’ve got three NethServer Community Edition installations: in my living room (thanks to @mywife), in the google cloud (thanks to @etino) and at my company’s biggest client. I’ve got a very specific use for them in mind, alert me 24/7 when a couple of business critical Sles HA clusters are going down, and a long term goal, make my client decide to replace Tivoli ITM and IBM Director with Nagios and OCS.
Fact is, despite being HA clusters, those Sles will go down on a non-regular basis due to mysterious san hiccups The San Team hasn’t still been able to trace, and when they do go down I might or might not receive notification from the ITM agents monitoring the servers,“why ?”,“we don’t know yet”, The ITM Team tells me: they’re working on it. I know they’re not, both Teams, so I adapted: in a couple of weeks I had my sles box with Omd monitoring the sles clusters, sending me out trusty sms messages, 5 seconds after the damage; and while I was at it, monitoring the other 300 servers on my watch, 60 Oracle database servers, web sites, backups, and the kitchen sink too, you know you can get lost browsing the list of available nagios checks.
24/7 means I can be out of office when shit happens, and even though I’ll receive the sms, I might wanna take a look at the Nagios console first, maybe it’s defcon 1, all servers down: I can’t vpn into my client’s network using the smartphone, but I can replicate the inside Sles box Omd checks status in realtime to an outside Omd box on Aruba’s cloud listening at port 80, and point my nagios android app to that site. Right now I use Check_Mk, its gui interface is better than the flat standard nagios one, but still you have to like it and it’s some way to go to fully automate the host add procedure. I had no plans for inventory, then I saw OCS at the Partners’ Meeting, and decided to give it a go.
My plan is to drop sles+omd and replicate the present scenario, inside server --> outside server --> smartphone, on NethServer + Adagios + OCS, adding some kind of push notification system (what’sapp like) to replace the sms system.
As for hardware:
Neth1, in the LivingRoom is on an dual proc Intel server built from scratch
Neth2: on Google, in a virtual machine in their cloud with the smallest hw configuration possibile to save the free 300$
Neth3: at the client’s site, is on a vSphere 6 farm made up of an IBM BladeCenter-E with 6 Ibm HS21 blades (they’re not on the vSphere 6 compatibility list, but they don’t know that and seem to work fine) and a fibre channel storage (IBM DS3400).