N-Central - Agents Disconnected

Looks like we're on the correct path. Gotta have the public face of it resolving properly. And yeah port 443 often already conflicts with something....tis where multiple IPs (a full block) comes into play to help you out.
And silly me thought, why would I need 5 static IP's when they offered me 1 or 5.........
 
This is why I left N-Able, agents were going offline all the time. At one point half my agents didn't work. Their support only came up said to reinstall the agent each time and as it was happening to some of the same clients multiple times I ended up leaving them. I had one of their engineers go over my setting, he made a few suggestions that I followed but it still happened.

Hope you get it sorted though, oh and mine was hosted by them.

Hi - what solution did you go with? I am on their hosted plan for ~2 weeks, and realizing the lack of functionality between the host & on-premise. Don't like that fact that I am finding out the difference post going live, but I understand some functionality not available on cloud. I did run into agent disconnect issues and didn't like the manual process of install/un-install that I had to go through.
 
I did run into agent disconnect issues and didn't like the manual process of install/un-install that I had to go through.

What type of client networks are you running into this on?
In some "takeover" situations we've had, where we take over a clients network...I've noted communication issues if DNS was not setup properly on the network. Create the domain admin account to give the N-Central probe and agents admin privvies, install the probe, push out/deploy agents...wham, done!
If DNS is not setup properly however, probe won't resolve the clients perfectly, and clients won't resolve the probe host perfectly.
Also check on firewall rules...not only in/out of the network, but on the Windows (or 3rd party if one) firewall for local communications. N-Able has guides on this.

We never tried it on a cloud host....as, from the beginning, we were going to host it in our server room. So we could control the speed/horsepower of the host server, and our firewall, and internet connection. We wanted the best possible remote performance, so we have a large block of public IPs with our fiber connection....our main office and e-mail spam filters and bench area run on a different router, our Untangle firewall. So those IPs get the traffic. Our N-Central server is behind a Ubiquiti EdgePro router..a wicked..wicked fast, ultra low latency router..and it's on a separate public IP. No competing traffic. We would have done this setup no matter what RMM we chose...it's just how we roll.
 
Hi - what solution did you go with? I am on their hosted plan for ~2 weeks, and realizing the lack of functionality between the host & on-premise. Don't like that fact that I am finding out the difference post going live, but I understand some functionality not available on cloud. I did run into agent disconnect issues and didn't like the manual process of install/un-install that I had to go through.
I ended up going with Logicnow who have been bought over by solarwinds
 
What type of client networks are you running into this on?

Smaller clients, Lawyer/Dental/Insurance offices. (5-10 nodes) It seems worse without a probe. Lawyer's office all Mac, no server. All fine last night @2:07am when I closed my session. Left my bash running cleaning up old backup-sets and this morning poof! No Mac in sight for me to wrap it up. now I am going onsite and pushing my schedule back. We will see if a network/power-failure issue, but doubt it since happening across the board.

Screen Shot 2016-09-29 at 7.47.02 AM.png

In some "takeover" situations we've had, where we take over a clients network...I've noted communication issues if DNS was not setup properly on the network. Create the domain admin account to give the N-Central probe and agents admin privvies, install the probe, push out/deploy agents...wham, done!
If DNS is not setup properly however, probe won't resolve the clients perfectly, and clients won't resolve the probe host perfectly.
Also check on firewall rules...not only in/out of the network, but on the Windows (or 3rd party if one) firewall for local communications. N-Able has guides on this.

One minute up, another minute off. It's Christmas. I am getting ready to fix an issue, but I can't get to it. I end up spending time going around to get on the device. an hr later after I am done, WALA it's online and healthy.

Different client - This one XP, but I have the same issues on Win7's - Working with the client on upgrading the clunkers.
Screen Shot 2016-09-29 at 7.48.28 AM.png

We never tried it on a cloud host....as, from the beginning, we were going to host it in our server room. So we could control the speed/horsepower of the host server, and our firewall, and internet connection. We wanted the best possible remote performance, so we have a large block of public IPs with our fiber connection....our main office and e-mail spam filters and bench area run on a different router, our Untangle firewall. So those IPs get the traffic. Our N-Central server is behind a Ubiquiti EdgePro router..a wicked..wicked fast, ultra low latency router..and it's on a separate public IP. No competing traffic. We would have done this setup no matter what RMM we chose...it's just how we roll.

awesome setup. Their Cloud solution flaky. I hope they work with me.

Thanks!
 
Smaller clients, Lawyer/Dental/Insurance offices. (5-10 nodes)
View attachment 6436

Hmm...so that example there, based on the host name, shows an ISP gateway being the DHCP/DNS....so it's probably a workgroup/peer to peer network huh?
I can't help but wonder if the....poor features, of the typical ISP supplied gateways, may be a common denominator for people that have issues such as this. Overworked port 443, or failure to resolve required n-central servers quick enough.
 
Back
Top