1

I'm the one-man IT department for a small business. We use Office 365 Exchange with client machines all running Windows 7 Pro, and either Outlook 2013 or 2016 on each machine. Occasionally, (ranging from 0 to 5 times per working day, per machine) the Outlook clients will lose connection to Office 365 whilst still remaining connected to, well, everything else.

While this is happening, nslookup for outlook.office365.com will fail on whichever machine is currently experiencing the connection issues. The DNS requests will time out. Other machines on the same network (and joined to the same domain, using the same DNS server [we're using active directory with a Server 2016 Essentials domain controller doing DNS locally]) will still be able to connect and nslookup will not fail on the other machines.

This usually lasts for 30 minutes to an hour before the connection is restored, or until the client machine is restarted, which instantly fixes the problem. Closing Outlook and re-opening does not fix the issue.

Any ideas here? I've been dealing with this for about a month now. Multiple support calls to Microsoft with about 10 hours total spent on the phone have gotten us nowhere on this issue. Support has screenshared with me while i'm remoted into the DC/DNS server. Still no fix.

Any ideas? I have tons more information/screenshots/debug outputs I can provide as needed.

user395062
  • 11
  • 1
  • 3
  • Are you using split-brain DNS for your public domain (define your public domain zone inside your internal DNS server for your LAN)? If so, I'd verify it matches your registrar's public domain zone file. – Jarrod L. J. Gibson Jan 13 '17 at 21:39
  • Nope, our public domain/website is being hosted externally. – user395062 Jan 13 '17 at 21:50
  • What kind of firewall do you use? – Jarrod L. J. Gibson Jan 13 '17 at 21:54
  • Just the built-in firewall on our Comcast Business gateway. We DID have a Cisco SA520 that died, which I did not replace. There's only about 20 devices on the network, total, so I figured I could get away with it. Please, please feel free to let me know if I'm being an idiot here. I have a vested financial interest in the company so I try to minimize our costs where I can, but if we need it, I'll get one. – user395062 Jan 13 '17 at 22:09
  • How are the DNS client settings on the DC and the clients configured? – joeqwerty Jan 13 '17 at 22:12
  • Clients are using the DC as their DNS server. I don't have any forwarders set up on the DC, just using root hints. The lookup zones are all basically the default entries from a fresh Server 2016 install as Domain Controller, plus reverse lookup zones for the client machines connected to the domain. Nothing special. – user395062 Jan 13 '17 at 22:22
  • Does the DC use only itself and 127.0.0.1 for DNS? You don't have your router or ISP DNS configured in the DNS client settings on the DC do you? – joeqwerty Jan 13 '17 at 22:28
  • I have the primary set as 127.0.0.1, I shouldn't have an alternate, correct? – user395062 Jan 13 '17 at 22:38
  • Try to use Wireshark to see what’s going on on the LAN when DNS requests does not work. We used to have malfunctioning router which was setting lot of RST packets randomly to the network and it caused intermittent issues. – Jozef Izso Jun 04 '20 at 23:00

1 Answers1

0

You should have your DC set to point to it's own static IP address on your LAN for DNS. I would also setup forwarders in DNS on your DC that point to your ISP's DNS servers.

NOW-Admin
  • 72
  • 3
  • I have the DC set up like that already, but I'll add the forwarders. I believe we were running that configuration previously, but it's not the case right now. – user395062 Jan 16 '17 at 19:24
  • Keep us posted on your progress. – NOW-Admin Jan 18 '17 at 19:32
  • I've added our ISP's (Comcast) DNS servers as forwarders. It's not made a difference. I'm still on a relentless google/technet search as I keep exploring what could possibly be causing this. I appreciate the help, guys, we're just not getting anywhere quite yet. – user395062 Jan 19 '17 at 16:43
  • What is your antivirus software that is installed on the client computers? Have you tried running with this and the related software firewall turned off? – NOW-Admin Jan 20 '17 at 04:37
  • I should probably add that this all started when I removed our malfunctioning Cisco SA520 hardware firewall. Before I removed it, I checked the ruleset and there were only 5 rules, none of which seemed to be relevant to what we're doing now. We had previously been running our own on-premises Exchange server, (which I eliminated 1.5 years ago in favor of Office 365) and all of the rules appeared to be related to that. Any idea what the SA520 could have been doing that would have been preventing this? I haven't set any special rules in the Comcast firewall we're using now. – user395062 Jan 20 '17 at 21:28
  • Also, below is the output of nslookup when the clients are experiencing connection issues. nslookup outlook.office365.com 192.168.0.20 Server: INS-SBS02.company.local Address: 192.168.0.20 DNS request timed out. timeout was 2 seconds. DNS request timed out *** Request to INS-SBS02.instrucon.local timed-out nslookup outlook.office365.com 75.75.75.75 Server: cdns01.comcast.net Address: 75.75.75.75 DNS request timed out. timeout was 2 seconds. DNS request timed out. timeout was 2 seconds. *** Request to cdns01.comcast.net timed-out – user395062 Jan 21 '17 at 01:05
  • This sounds like a loss of response from the Comcast DNS server. Is your Comcast modem/router on UPS power? I am wondering if it has an intermittent issue due to some power surge. See if you can ping the IP address of the DNS server from Comcast on the computer when it cannot reach the 365 server. If the Comcast device is dropping intermittently, it will need replaced. If you are unable to ping the DNS server, you might need to change your forwarders. – NOW-Admin Jan 22 '17 at 00:28
  • The Comcast gateway is on a UPS in the server room; every other machine has no issues. This usually happens to only one or two client machines at a time. Everything else on the network will ping/nslookup/etc Office 365 no problem, including our domain controller. – user395062 Jan 23 '17 at 22:04
  • Are the computers dropping connection random or always the same ones? – NOW-Admin Jan 24 '17 at 18:09
  • All of the machines using Outlook are affected, just not at the same time. – user395062 Jan 24 '17 at 18:44