So this is a very perplexing situation. We employ a Linux server to run an ancient ERP (accounting and sales orders) and it is disconnecting from the local network at seemingly random intervals. I'm going to list the entire ordeal just in case there is something important that I'm missing.
Two weeks ago Friday, I decided to remove my local peer to peer network from a shared IP phone network, because I didn't realize that a managed phone network also meant locking me out of my local peer to peer. So I disconnected all the devices from the managed network and created a new peer to peer with a new IP, modem, switch and router, completely separate. I'm not a network whiz, so it took me a while to get all the bells and whistles working properly, but I had everything working within 24 hours.
While working remotely (using SSH/Telnet) from home on Sunday, I noticed that the email server wasn't working (part of the Linux server). I found it strange that this coincided with my network switch over, but since everything else was working fine, I proceeded to raise a trouble ticket with my hardware vendor citing email failures from the server.
Monday morning, support responds that there was nothing from what they could see, no errors, no problems. But still, no email. So I reboot the server and voila, emails are working again. So a few days pass and all is well, until Friday when we start getting local network disconnects from the server. We can tell immediately since our PC terminal emulators crash. So I go back to my IT closet and do my best Sherlock Holmes impersonation and end up shrugging my shoulders and just...reboot. Working, for now.
Saturday morning I wake up to no remote connection. I drive to the office and find that the entire rack is shutdown. Now I have a new UPS and never connected it to the server. Nothing else in the office is off, just the rack. So I check the UPS and it's just off, completely off. Now I actually don't think this is related, because there is no data connection from the UPS to anything else in the network, but it's a bizarre occurrence that I figured I would mention in the event that someone recognizes it as a symptom of something else. So I boot up the server and all is well for the rest of the day. Connection is stable, no problems. Power has been stable for 3 days and counting as of this post.
Sunday morning comes and my Linux login is going very slow, like a 10 second delay between username and password. My stomach sinks.... So I remote desktop into the office to check some settings and to my utter amazement, I can't access the server from the local network. I can login remotely (SSH/Telnet), but the local network connection to the server is DEAD.
I arrive this past Monday to find nothing changed. I still cannot access the server via the local network. No ping, no http, no terminal emulator....nothing. So I start digging into the router and I find that the server is listed in the connected devices. So the router sees the server, but it isn't routing the connection locally, only remotely. I feel like this is a major clue, but I haven't any idea why.
I end up breaking down and calling NetGear tech support. Let's just say that was completely and utterly useless. So I decided to make some changes myself, just to eliminate possible weak links. I bypassed the switch and connected the server directly to the router. I reserved the local address for the server to avoid ip conflict. I even eliminated DHCP from assigning any ip address in that range, still nothing helps.
One thing I noticed is that connection seems to drop during times of inactivity for an hour or more. I can't be certain if this is correlation or causation, but it's something that has me curious. My first thought was that the network card was going to sleep, but the server doesn't have any power management software, and of course that wouldn't support the fact that it's still accessible remotely. So I'm still leaning towards this being a router issue, as that's the only piece of new hardware left that hasn't been bypassed or replaced. Everything else was working prior to the switchover.
Some are going to suggest that I replace the router. That is certainly my last option, but in my defense, the router is brand new and while that doesn't preclude it from being defective, it seems like a really unlikely scenario. But of course, this entire problem seemed really unlikely too.... I did have an older Linksys router that I tried to use in place of the Netgear, and that thing was hopelessly shitty. I couldn't even connect to it half the time during setup. I'm not here to trash Linksys, but damn Belkin....
In conclusion, the server is dropping connection to the router (or vice versa) on a regular basis, but only on the local network, remote access via SSH/Telnet functions, albeit with some odd behaviors. The email server does not work during these outages at all. Rebooting the server is the only thing that solves the issue at the present time.
Here's some details of my setup
Router: Brand New NetGear Nighthawk AC1750, R6700v3, Port 22 forwarded per vendor software requirement Switch: JGS524 IP: Frontier FIOS, Static IP Network: Peer to Peer, Windows 10, 7 and Linux Redhat Server
If you have read this far, thank you for taking interest in one mans problems. I appreciate anything you could contribute to this issue.
