Site to site slowdowns after VLAN additions

[brad-denham]

This past Saturday I reconfigured one of our remote sites networks to add VLAN?s. The site has a 3620, 4 - 3524 powered switches (users) and one 2970G-24TS (servers and power users). Once the configuration had been completed, I tested the new VLAN?s ability to get the correct IP addresses (DHCP) and if I could access remote servers and the Internet. All was working well. Approximately an hour after I had left the site I received a call that simple drag and drop (Windows and OS X) file transfers were extremely slow. Upon returning to the site I connected my laptop into one of the switches and verified this behavior. A quick scan of Ethernet and serial interfaces on the router showed no problems, sh proc cpu looked fine, there was nothing glaring as to what was wrong. All of the trunk interfaces between the switches are all set to 100 full, so no duplex or speed mismatches there. The connection from the first switch to the routers connection was checked for the same thing and all was correctly configured for the same. So I rebooted the router and the switches and all symptoms of the slow network disappeared.

Things were fine for the rest of Saturday, Sunday, and yesterday. This morning I received a call and the slow network was back, I headed over to the site, plugged my laptop into the network and things had indeed become extremely slow. I went through the steps mentioned above with the exception or rebooting all network devices. This time I just rebooted the 3620 and after it came back up the normal network speed had been restored.

During the times of extreme slowness the LAN was not affected, they have local servers connected to the 2970G. Users on the 3524?s were able to transfer files at normal speed. Just the interaction with servers (Exchange, FTP, Domain Controllers) across the WAN was negatively affected.

I am attaching 3 files, one is a sho ver, and the other two are the 3620 config before the VLAN?s and after. I am not sure what I am missing here, if it is something in my config, or if this is possibly just a weird fluke. There were no problems with slow network access before the reconfigure.

Thanks for taking a look.

[Hieu Cao]

Brad,

I think there's a problem with your "rate limit" statement. Burst-normal and Burst-max values are the same (128000 128000 128000). Try burst-normal=4000, burst-max=8000. Or you can try taking out the both the Rate Limit statements for now and monitor the router to see if it matters.

You might want to consider prioritize critical network traffic using QoS. Also, no speed/duplex was configured for interface FastEthernet0/0.1

Good luck

HTH,

hieu

[brad-denham]

I have removed the rate-limit statement (I also modified my fat fingering of it as well) and I still have the same problem. The rate-limiting statement is on the subinterface that only guests access. The reason for the rate limiting is so that if a guest connects they cannot monopolize the bandwidth.

[Hieu Cao]

Can you try removing the "ip access group in" under both sub-interfaces. You didn't have them in your original config.

hieu

[glen.grant]

I guess I would start taking things off one at a time that weren't there before , like ip route chache flow off of E0/0 , then the ACL's one at a time off the subinterfaces .

[brad-denham]

Thank you very much for the suggestions.