kstak Posted March 26, 2013 Posted March 26, 2013 First post but this appears to be a great forum so I hope to participate more... Here is my problem: I have two procurve 4108GLs that are connected by fiber. They both have the latest firmware. One of them (A) has 2 vlans. Vlan1 is the default with a subnet of 192.168.2.0/24. Vlan2 has a subnet of 192.168.1.0/24. The switch is routing between the vlans successfully. Switch B has the default vlan1 with a subnet of 192.168.2.0/24. The switches are connected via the fiber on subnet 192.168.1.0/24. There is no other configuration. Everything appears to work fine. However, under high load I start loosing packets going from vlan1 on switch B (192.168.1.0/24) to vlan1 on switch A (192.168.2.0/24). I do not loose packest on swtich A going from vlan1 to vlan2. I do not loose packets going from vlan1 switch B to vlan2 switch A. The vlans shouldn't matter since switch B is basically an expensive unmanaged switch. Can the routing on switch A just be overloaded? I see ~40% cpu average. But why does switch A vlan1 to vlan2 have no issues then? Do the vlans matter even if I am not passing any info? Any insight would be greatly appreciated.
Marshall_IT Posted March 30, 2013 Posted March 30, 2013 I take it the first switch is the default gateway for both vlans. Also is vlan1 on both switches tagged with the same number? The link between switch 1 and 2 is set to untagged on vlan1?
CyberNerd Posted March 31, 2013 Posted March 31, 2013 (edited) my top tips for debugging network loops: on core switch: ping offending device “display arp” to determine mac address “display mac-address xxxxxxxx” to determine port or aggregation group ping offending device when it goes offline “display mac-address xxxxxxxx” to determine port or aggregation group note change of port. repeat on the other switches to determine loop. fix loop or shutdown port Edited March 31, 2013 by CyberNerd
kstak Posted April 1, 2013 Author Posted April 1, 2013 Disclaimer: This isn't my configuration. I inherited this. I personally would change a number of things including taking the routing off the switches but I have to prove that will fix it.... also I changed these ips so if I missed something, apologies. I just noticed that switch A has "ip default-gateway 192.168.1.7" which is wrong but "sh ip route" and the ip routing configuration have the proper gateway of ip default-gateway 192.168.2.4. Switch A won't let me take out "ip default-gateway" without disabling ip routing first and I can/t do that right now. I am not sure that would cause an issue because the routing table appears correct but it will be my next step just to see if it helps. Today I am not seeing much loss (1 dropped in 1000) but still do see some lag spikes. It is only from some on switch A going to some device on switch B and it is only from 192.168.2.0/24 to 192.168.1.0/24. Swtich B is basically un-configured. Switch A is semi-configured. The client has switch A vlan2 192.168.1.7 as the main router for their 192.168.1.0/24 network. There are some servers on 192.168.2.0/24. The main internet router/firewall is at 192.168.1.4 @Marshall_IT A1 on both switches is the link. They are not tagged. @CyberNerd I am pretty confident there isn't any looping. Switch B is all servers with short patch cables directly to the servers. I suppose something could be looped on A but I would think I would be seeing other issues on A. Switch A ; J4865A Configuration Editor; Created on release #G.07.117 hostname "HP ProCurve Switch 4108GL" cdp run qos-passthrough-mode module 5 type J4863A module 2 type J4862B module 3 type J4862B module 4 type J4862B module 1 type J4864A module 8 type J4862B module 6 type J4863A interface E1 no lacp exit interface E2 no lacp exit interface E3 no lacp exit interface E4 no lacp exit interface E5 no lacp exit interface E6 no lacp exit interface B1 no lacp exit interface B2 no lacp exit interface B3 no lacp exit interface B4 no lacp exit interface B5 no lacp exit interface B6 no lacp exit interface B7 no lacp exit interface B8 no lacp exit interface B9 no lacp exit interface B10 no lacp exit interface B11 no lacp exit interface B12 no lacp exit interface B13 no lacp exit interface B14 no lacp exit interface B15 no lacp exit interface B16 no lacp exit interface B17 no lacp exit interface B18 no lacp exit interface B19 no lacp exit interface B20 no lacp exit interface B21 no lacp exit interface B22 no lacp exit interface B23 no lacp exit interface B24 no lacp exit interface C1 no lacp exit interface C2 no lacp exit interface C3 no lacp exit interface C4 no lacp exit interface C5 no lacp exit interface C6 no lacp exit interface C7 no lacp exit interface C8 no lacp exit interface C9 no lacp exit interface C10 no lacp exit interface C11 no lacp exit interface C12 no lacp exit interface C13 no lacp exit interface C14 no lacp exit interface C15 no lacp exit interface C16 no lacp exit interface C17 no lacp exit interface C18 no lacp exit interface C19 no lacp exit interface C20 no lacp exit interface C21 no lacp exit interface C22 no lacp exit interface C23 no lacp exit interface C24 no lacp exit interface D1 no lacp exit interface D2 no lacp exit interface D3 no lacp exit interface D4 no lacp exit interface D5 no lacp exit interface D6 no lacp exit interface D7 no lacp exit interface D8 no lacp exit interface D9 no lacp exit interface D10 no lacp exit interface D11 no lacp exit interface D12 no lacp exit interface D13 no lacp exit interface D14 no lacp exit interface D15 no lacp exit interface D16 no lacp exit interface D17 no lacp exit interface D18 no lacp exit interface D19 no lacp exit interface D20 no lacp exit interface D21 no lacp exit interface D22 no lacp exit interface D23 no lacp exit interface D24 no lacp exit interface A1 speed-duplex 1000-full no lacp exit interface A2 no lacp exit interface A3 no lacp exit interface H1 no lacp exit interface H2 no lacp exit interface H3 no lacp exit interface H4 no lacp exit interface H5 no lacp exit interface H6 no lacp exit interface H7 no lacp exit interface H8 no lacp exit interface H9 no lacp exit interface H10 no lacp exit interface H11 no lacp exit interface H12 no lacp exit interface H13 no lacp exit interface H14 no lacp exit interface H15 no lacp exit interface H16 no lacp exit interface H17 no lacp exit interface H18 no lacp exit interface H19 no lacp exit interface H20 no lacp exit interface H21 no lacp exit interface H22 no lacp exit interface H23 no lacp exit interface H24 no lacp exit interface F1 no lacp exit interface F2 no lacp exit interface F3 no lacp exit interface F4 no lacp exit interface F5 no lacp exit interface F6 no lacp exit ip default-gateway 192.168.1.7 ip routing vlan 1 name "DEFAULT_VLAN" untagged A2-A3,C1-C24,H1-H24 ip address 192.168.2.2 255.255.255.0 no untagged A1,B1-B24,D1-D24,E1-E6,F1-F6 ip igmp exit vlan 2 name "VLAN2" untagged A1,B1-B24,D1-D24,E1-E6,F1-F6 ip address 192.168.1.7 255.255.255.0 exit ip route 0.0.0.0 0.0.0.0 192.168.2.4 ip route 10.0.2.0 0.0.0.0 192.168.1.5 no stack password manager Switch B Running configuration: ; J4865A Configuration Editor; Created on release #G.07.117 hostname "HP ProCurve Switch 4108GL" cdp run qos-passthrough-mode module 1 type J4864A module 5 type J4863A module 2 type J4862B module 3 type J4862B module 4 type J4863A module 7 type J4863A module 8 type J4863A module 6 type J4863A interface A1 speed-duplex 1000-full no lacp exit interface A2 no lacp exit interface A3 no lacp exit interface E1 no lacp exit interface E2 no lacp exit interface E3 no lacp exit interface E4 no lacp exit interface E5 no lacp exit interface E6 no lacp exit interface B1 no lacp exit interface B2 no lacp exit interface B3 no lacp exit interface B4 no lacp exit interface B5 no lacp exit interface B6 no lacp exit interface B7 no lacp exit interface B8 no lacp exit interface B9 no lacp exit interface B10 no lacp exit interface B11 no lacp exit interface B12 no lacp exit interface B13 no lacp exit interface B14 no lacp exit interface B15 no lacp exit interface B16 no lacp exit interface B17 no lacp exit interface B18 no lacp exit interface B19 no lacp exit interface B20 no lacp exit interface B21 no lacp exit interface B22 no lacp exit interface B23 no lacp exit interface B24 no lacp exit interface C1 no lacp exit interface C2 no lacp exit interface C3 no lacp exit interface C4 no lacp exit interface C5 no lacp exit interface C6 no lacp exit interface C7 no lacp exit interface C8 no lacp exit interface C9 no lacp exit interface C10 no lacp exit interface C11 no lacp exit interface C12 no lacp exit interface C13 no lacp exit interface C14 no lacp exit interface C15 no lacp exit interface C16 no lacp exit interface C17 no lacp exit interface C18 no lacp exit interface C19 no lacp exit interface C20 no lacp exit interface C21 no lacp exit interface C22 no lacp exit interface C23 no lacp exit interface C24 no lacp exit interface D1 no lacp exit interface D2 no lacp exit interface D3 no lacp exit interface D4 no lacp exit interface D5 no lacp exit interface D6 no lacp exit interface G1 no lacp exit interface G2 no lacp exit interface G3 no lacp exit interface G4 no lacp exit interface G5 no lacp exit interface G6 no lacp exit interface H1 no lacp exit interface H2 no lacp exit interface H3 no lacp exit interface H4 no lacp exit interface H5 no lacp exit interface H6 no lacp exit interface F1 no lacp exit interface F2 no lacp exit interface F3 no lacp exit interface F4 no lacp exit interface F5 no lacp exit interface F6 no lacp exit ip default-gateway 192.168.2.7 vlan 1 name "DEFAULT_VLAN" untagged A1-A3,B1-B24,C1-C24,D1-D6,E1-E6,F1-F6,G1-G6,H1-H6 ip address 192.168.2.242 255.255.255.0 exit
DMcCoy Posted April 1, 2013 Posted April 1, 2013 Could it not be because you are trying to pass a lot of traffic over a single 1Gb link? Which would not be surprising under load.
kstak Posted April 1, 2013 Author Posted April 1, 2013 Could it not be because you are trying to pass a lot of traffic over a single 1Gb link? Which would not be surprising under load. I don't think bandwidth is the issue. The link mostly handles low bandwidth traffic like rdp and ssh but it does handle a good deal of that traffic. It could be a packet per second issue but think the HP 4018GL is rated 72 million pps. If that was the issue, I think I would see the issues going from switch A to switch B w/o routing and I only see it from switch A to B when it is first routed. Could the routing be handled differently if the traffic doesn't originate from the switch? That seems like a long shot but I plan to do some testing to try and verify that but have not had the time yet. I also plan to test a secondary router when they let me. More pieces to the puzzle... switch A port A1 counters - Up 9 days. Status and Counters - Port Counters for port A1 Name : Link Status : Up Bytes Rx : 1,402,845,252 Bytes Tx : 2,159,397,686 Unicast Rx : 3,078,544,823 Unicast Tx : 3,479,354,851 Bcast/Mcast Rx : 4,054,308 Bcast/Mcast Tx : 2,317,403 FCS Rx : 0 Drops Rx : 3351 Alignment Rx : 0 Collisions Tx : 0 Runts Rx : 0 Late Colln Tx : 0 Giants Rx : 0 Excessive Colln : 0 Total Rx Errors : 0 Deferred Tx : 0 switch B port A1 counters Status and Counters - Port Counters for port A1 Name : Link Status : Up Bytes Rx : 2,324,658,228 Bytes Tx : 1,558,440,133 Unicast Rx : 3,479,984,773 Unicast Tx : 3,079,287,731 Bcast/Mcast Rx : 2,319,631 Bcast/Mcast Tx : 4,058,700 FCS Rx : 0 Drops Rx : 0 Alignment Rx : 0 Collisions Tx : 0 Runts Rx : 0 Late Colln Tx : 0 Giants Rx : 0 Excessive Colln : 0 Total Rx Errors : 0 Deferred Tx : 0
_Adam_ Posted April 3, 2013 Posted April 3, 2013 Your explanation of the setup is full of inconsistencies, however based on your post #7 I would either suggest using another patch lead or set the interfaces back to auto as a test and see what they negotiate themselves to, which could show an issue if it's not 1000/full. Although your config could be better (not your fault as you say) nothing in it should cause the drops.
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