smacrory1980 Posted February 25, 2015 Posted February 25, 2015 Hi We seem to have some general network issues with the main symptoms being Slow internet, pages taking forever to load etc, when used by a number of pupils. Ping tests on the LAN regularly showing a time of over 100ms There doesn't seem to be an issue accessing shares or logging on unless a number of laptops are used wirelessly in one area, which is to be expected I suppose As for the slow internet the council maintain that there are peaks in bandwidth usage due to unwanted traffic from our LAN. As a result I carried out a wireshark test which showed a high percentage of the traffic was Netbios, dropbox and IPv6 traffic. I have tried disabling these settings and have modified the dropbox settings. The problems still exist. Other things I have tried inlcude with/without STP, updating NIC drivers, checking cabling etc. One area that I could do with help on is DNS as I suspect it may not be configured correctly as the wireshark logs seem to indicate that the clients are going straight to the secondary DNS, is that ok? It is set up as follows DHCP Server 'Default Gateway (Council Router)' DNS Servers 'Local DC' and 'Council proxy' Also, is there anything that I should check on the server with regards to this? It was all set up before I started with no documentation so any ideas or things I could check are much appreciated. We have 1 server, server 2008r2, and clients on win7 Cheers Steve
supawat1 Posted February 25, 2015 Posted February 25, 2015 I had an issue where some web pages could loaded up fine and some dontv loaded up at all. It turned out to be MTU setting within the router as it was set too low. May be you could check that.
Ephelyon Posted February 25, 2015 Posted February 25, 2015 What are ping times between switches like? That could help isolate the issue to the network generally or certain nodes specifically.
localzuk Posted February 25, 2015 Posted February 25, 2015 Does the council DNS handle any DNS for your internal network? If not, it should not be used by your client PCs. Instead, your onsite DNS should be set up to forward DNS requests to the council DNS. That way, all requests get sent internally, and the DNS server forwards them up and out of your network if it cannot fulfil them itself. With your ping tests, is that pinging DNS, so there's some name resolution, or is that pining an IP address on your network?
machin05 Posted February 25, 2015 Posted February 25, 2015 Does the council DNS handle any DNS for your internal network? If not, it should not be used by your client PCs. Instead, your onsite DNS should be set up to forward DNS requests to the council DNS. That way, all requests get sent internally, and the DNS server forwards them up and out of your network if it cannot fulfil them itself. With your ping tests, is that pinging DNS, so there's some name resolution, or is that pining an IP address on your network? Was just about to suggest that
smacrory1980 Posted February 25, 2015 Author Posted February 25, 2015 The pings seem to be the same whether DNS or IP. The councils router handles DHCP and provides clients with our local server as the primary DNS and their server as the secondary DNS. Our server is also set up to forward to their DNS. So your saying that just the primary DNS should be dished out to clients? Some of the clients appear to be going directly to the council DNS so I'm unsure how it would affect them if I turned the secondary DNS off.
Ephelyon Posted February 25, 2015 Posted February 25, 2015 Again, what happens if you ping one switch from another switch?
machin05 Posted February 25, 2015 Posted February 25, 2015 Yeah, your clients should only be using your internal DNS. Then set up your DNS server with a Forwarder to the council one so that any requests for anything outside of your local network will be handled by your internal DNS server
smacrory1980 Posted February 25, 2015 Author Posted February 25, 2015 Pings between switches are usually 5 or 6ms. I got a couple that went to 77ms and 13ms, but when I tried again they were at 4/5/6ms. These are 1GB switches so guessing they should really be ,<1ms. Any ideas?
Ephelyon Posted February 25, 2015 Posted February 25, 2015 Good shout, that. Also check that one of your core switches is actually the root of the spanning tree ("show spanning-tree" on ProCurve switches); we had an issue once where County's Avaya IP telephony gateway had somehow decided to be the root, which caused a bit of confusion for the switches...
smacrory1980 Posted February 25, 2015 Author Posted February 25, 2015 I had the network install chaps come in to check for loops and they couldn't find any which is what prompted me to go down the road of looking into any windows or DNS settings. I'll have a check for the root and any other errors, thanks
achedgy Posted February 25, 2015 Posted February 25, 2015 Do you have ipads? Bonjour traffic can grind things to a halt, found isolating them onto a guest wireless network improved network performance massively...
smacrory1980 Posted February 25, 2015 Author Posted February 25, 2015 Yes we have ipads and all the teachers laptops have itunes installed so that could well be having an affect, thanks
Jamo Posted February 25, 2015 Posted February 25, 2015 That sounds an awful lot like network issues to me. Wireshark will only show you traffic which hits your network interface, so that's broadcast traffic, or traffic directed at your IP/MAC. It might be worth looking at the switches and checking the counters, on a cisco its show int gi0/0 haven't used HP for a while but it will be a similar command. Look for Errors/Discards, if there are loads, clear the counters and leave it for a few days and have another look. Your errors and discards should be low, even on aggregate/portchannel links. I agree with the others, with an active directory network you shouldn't target an outside DNS server (unless you are testing something). It won't break anything its just not recommended due to AD being tied so closely to DNS. However I don't believe that DNS is the root cause of the issue you mentioned.
smacrory1980 Posted February 26, 2015 Author Posted February 26, 2015 Think I'll give this a go, found on another thread... Spanning Tree will not prevent all kinds of loops and that isn't its purpose in the first place; what you want is Loop Protect. You can enable it on the Procurves by connecting via the cli and executing the following commands: loop-protect all loop-protect trap loop-detected loop-protect disable-timer 60 write mem (if you want to save this config permanently) this will disable any looped ports for 60 seconds and generate an snmp trap.
smacrory1980 Posted March 11, 2015 Author Posted March 11, 2015 Hi This was eventually traced back to the local councils DHCP server dishing out the incorrect DNS suffix. Thanks everyone for your assistance and ideas.
Recommended Posts
Create an account or sign in to comment
You need to be a member in order to leave a comment
Create an account
Sign up for a new account in our community. It's easy!
Register a new accountSign in
Already have an account? Sign in here.
Sign In Now