-1

I have users who complain of network share disconnections happening in the background and it causing issues to their applications which are accessing these network resources.

Because of this I decided to enable the other NIC on this server (HP Proliant DL 320 G5p). The first time I did this many problems happened with DNS host resolution for external IPs and I had to back out the change. Both NICs will talk to the same internal network and one will go to each of my switches.

Today I decided to try this again and I wanted to make sure I set up the secondary connection correctly. When setting the IP I did not enter a gateway or any DNS server values. Under advanced on the DNS tab I unchecked the "Register this connections address in DNS".

I went into DNS manager and deleted my entries for this new IP that were made. Also did this on my secondary DNS server.

Do I have to define a persistent route as well? Some articles I have read today reference that but i'm not sure I need it since my NIC networks are common to each other.

edit: additional troubleshooting information as requested. This has happened intermittently for over a year so here are the most recent things I have tried over the last month or two.

  1. NIC driver updates applied
  2. nic cabling to upstream switch replaced
  3. connected nic to switch #2 instead of switch #1 and moved switch 2 as entry point to lan from firewall.
  4. At server - Power options to "turn off nic to save power" were disabled.
  5. On client workstations power settings enabled for full performance/always on. Nic power management disabled.
  6. Group Policy change to disable slow link mode to affect everyone.
  7. As ram resources constantly were running 75-85% I made a change to sql server to limit the amount of consumable memory. This was for a x86 WS2k8 w/ SP2 box. The other box has plenty of ram and didn't get this change.
  8. Changed binding order of NICs on the server with multiple cables.

No link-state related eventlog entries for system log. Nothing but bad packet warnings in dns event log. Not sure where else I should be looking. Teaming the nic isn't my desired result here. It's to round robin load balance. The servers are holding too many network roles as is and i've asked for a dedicated fileserver but not gotten it so i'm just trying to make do as best I can here.

Improve this question
8
  • 2
    There's nothing quite like having a problem and going completely off the reservation in trying to troubleshoot and resolve it. Do you normally just start "hacking" away when there's a problem? If you're trying to create a fault tolerant network adapter team you're doing it wrong. Even if you succeed in creating a fault tolerant network adapter team it's not likely to have any bearing on the problem. The problem is not likely due to the current network adapter losing connectivity. Maybe you should do some troubleshooting to better understand the problem before implementing a solution.
    – joeqwerty
    Dec 16, 2013 at 22:11
  • 1
    ^^^ Everything joe said. You haven't told us why you think that's the solution, and what troubleshooting you did to lead you to that conclusion.
    – mfinni
    Dec 16, 2013 at 22:15
  • Oh, screw PC. This is like waving a auto-repeating belt-fed water-cooled multi-barrel shotgun wildly around the farm to down a sparrow that's hardly moving on the other side of the field.
    – Roman
    Dec 16, 2013 at 23:12
  • @Roman - do you care to comment on additional troubleshooting I should have done? I don't do this sort of stuff every single day and i'm still trying to learn here.
    – TWood
    Dec 17, 2013 at 14:15
  • 1
    Dear merciful Jeebus. DHCP and DNS, for a small office environment, put very little load on a machine, let alone choke the bandwidth available. That's not "a lot of roles", and "a lot of roles" generally means squat. Measure and test, don't guess.
    – mfinni
    Dec 18, 2013 at 4:30

2 Answers 2

Reset to default
2

There's a great article on TechNet on how to configure multiple NICs on a File Server running Windows 2008/2008 R2. You can find that article here.

However, I think what you're looking for is:

2.3 Standalone File Server, 2 NICs on server, single subnet enter image description here

You will need to configure both NICs with different IPs on the same network. You will also want to configure the DNS information, default gateway etc. If this is configured correctly, your clients should hit the server in a round robin fashion - this should spread the load somewhat over both NICs.

You can also do NIC Teaming/Link Aggregation. This will require you to install software from your hardware vendor as Windows Server 2008/2008 R2 do not do this natively.

enter image description here

However, you really should figure out why you're having issues before doing any of these. Check logs, check performance, check client machines, etc. Neither of these solutions will help you if it's a switch/networking issue.

Improve this answer
2
  • Cole, I appreciate the link. I don't know what i'm looking for other than the items I listed above in troubleshooting steps.
    – TWood
    Dec 17, 2013 at 15:40
  • Look at the caveats for the first diagram, which is what TWood was trying : It's important to note that applications other the file server and file client might not behave properly with this configuration (they might not listen on all interfaces, for instance). You might also run into issues updating routing tables, especially in the case of a failure of a NIC or removal of a cable. These issues are documented in KB article 175767. For these reasons, many will not recommend this specific setup with a single subnet.
    – mfinni
    Dec 18, 2013 at 4:32
1

As these guys stated I don't need this setup for my small network after all.

After further troubleshooting the actual cause of the issue that was leading me to want to multi-home my fileserver was a failing switch and a issue with my failover dns server. Once I replaced the switch and disabled failover dns all of my network share disconnection issues disappeared and the entire network responsiveness changed a great deal.

With my switch being a HP 1410-24g unmanaged it had nothing I could really check or reconfigure to help rule out a problem with it. I figured I could snoop the traffic with Wireshark and look for odd behavior, but I didn't know what I was looking for.

I do intend to team the nics on my servers sometime down the road once I know for sure that my issue here disappeared completely.

Thanks for the assistance!

edit: 1/10/14 - I was completely incorrect here. My entire problem is with the AV package that I am running on the servers (fep 2010). Even though I created role-specific policies for each server it would appear that the real time protection is what actually randomly disconnects my users. The server logs reflect dropped and malformed packets as well as blocked ones. I wasted my money buying a new switch, but now I at least have a spare for future uses.

Improve this answer
1
  • Sounds about right. Glad you found it without a lot more headache.
    – mfinni
    Dec 19, 2013 at 20:47

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .