localzuk Posted April 4, 2011 Posted April 4, 2011 I've got an issue connecting to our Cachepilot box which is connected to one of our edge switches, and set to be connected to our server vlan. Machines on all our edge switches can connect to it fine, in any of the different VLANs in place. Other servers can connect to it fine, and they're connected to the core switch and in the server vlan, the same as the cachepilot. No pinging, no traceroute. Nowt. If I put my desktop machine in the servers vlan, it works fine. The subnet masks are all correct too, so it isn't that. I'm at a bit of a loss! Any ideas?
Admiral208 Posted April 4, 2011 Posted April 4, 2011 I cant really see what the problem is here? Do you mean that its just your machine that cant connect?
localzuk Posted April 4, 2011 Author Posted April 4, 2011 All machines connected to our core switch and aren't in the server vlan cannot connect. But all machines on other switches, on othert VLANs can connect. So, the core switch is routing between the vlans for this unit for edge switches, but not for itself.
CyberNerd Posted April 4, 2011 Posted April 4, 2011 what can't connect? The clients? Possibly a windows firewall misconfig? Windows Firewall > exceptions > Custom list of domains. You can set in Group Policy..
localzuk Posted April 4, 2011 Author Posted April 4, 2011 All clients on the core switch, that aren't in the server vlan. No firewalls are enabled.
mac_shinobi Posted April 4, 2011 Posted April 4, 2011 When you say connect you mean as in a WEB GUI or as in just being able to ping / traceroute etc ? Only asking as wondering if proxy settings has anything to do with it ? It's not just the port on the switch you have your machine connected to thats duff or maybe mis configured ??
localzuk Posted April 4, 2011 Author Posted April 4, 2011 I mean all connectivity. Not just web, so proxy has nothing to do with it. And the ports are fine on the switch, as it'd mean 40+ of them were duff, on multiple modules. So, very unlikely.
AJT1 Posted April 4, 2011 Posted April 4, 2011 How are you routing between the VLAN's? Do you have an ACL somewhere only defining inter-vlan communication for specfiic protocols/IP's to route to the server VLAN?
localzuk Posted April 4, 2011 Author Posted April 4, 2011 How are you routing between the VLAN's? Do you have an ACL somewhere only defining inter-vlan communication for specfiic protocols/IP's to route to the server VLAN? Inter-Vlan routing is done by our core switch, a HP Procurve 5406zl. There are no ACLs in place, other than some rules to enable some broadcast traffic to traverse the VLANs (for ABTutor Control). So, all inter-vlan traffic is permitted.
SYNACK Posted April 4, 2011 Posted April 4, 2011 Sounds like your VLAN config on the core switch, are the workstations on the core in the same VLAN as the clients on the edge switches, if not are they setup into another (non-default) VLAN with the same routing rules?
localzuk Posted April 5, 2011 Author Posted April 5, 2011 (edited) Sounds like your VLAN config on the core switch, are the workstations on the core in the same VLAN as the clients on the edge switches, if not are they setup into another (non-default) VLAN with the same routing rules? Yup, same VLAN. The core vlan tables are the same as the edge switches. Everything else works across them, just not this one cachepilot box. All vlans are basically identical, just with different names and different IP address ranges. I've attached the core config and the edge config so you can see. I can't see anything that could cause this. Edit: the cachepilot is on port 24 of the edge switch there, and my machines are on C23 and D11 of the core.core.txtedge.txt Edited April 5, 2011 by localzuk
SYNACK Posted April 5, 2011 Posted April 5, 2011 Not sure about the config as I am more used to cisco, where are the ports assigned to the VLANs. Anyhow I was just wondering how the routing back from the cachpiolet is sorted. Is it a bunch of static routes or are you grabbing advertised routing information from the core using something like RIPv2. Have you tried creating another VLAN for the systems on the core switch with a different address space to see if that allows routing to take place. I guess the other thing to check is the firmware level on the core, maybe a bug in the core routing that could have been fixed in a later revision although I expect that you probably have the latest firmware on there anyhow.
localzuk Posted April 5, 2011 Author Posted April 5, 2011 I'm not sure what you mean regarding the routing. It is simply a normal 'computer' so nothing special happens routing-wise. I've not tried anything which requires a restart of the core yet - Ofsted in today, so not willing to fiddle... The core is on the latest firmware. Edge switches aren't.
SYNACK Posted April 5, 2011 Posted April 5, 2011 The cachepiolet connects to the internet right and acts as a proxy? If it has a default route that points out to the interent it means that if you do any internal routing (which you do) it means that the cach piolet is not diretly connected to all the internal subnets and so does not know about them, in order to respond to them it would try to send the reply via its default gateway (the internet). The way around this is to tell the proxy about the internal networks and give it a route to get to them (the internal core router address). You can do this manually via static routes on the box or automagicly by advertising the routes from the core via a routing protocol like RIPv2(need v2 as you are using CIDR). If one of these static routes in the proxy was a little screwy then it could cause issues like this. I may have misunderstood the situation so could be off on a tangent though.
localzuk Posted April 5, 2011 Author Posted April 5, 2011 No, we don't use it as a proxy. It is used for hosting Channel 4 Learning stuff (clipbank and espresso). So, it is a web server basically, which automatically pulls data from them and caches it.
SYNACK Posted April 5, 2011 Posted April 5, 2011 No, we don't use it as a proxy. It is used for hosting Channel 4 Learning stuff (clipbank and espresso). So, it is a web server basically, which automatically pulls data from them and caches it. Ah, I had got the wrong impression of what it did, disregard that route distribution stuff as it does not apply. I would be looking to a switch restart and/or creating another access VLAN for testing on the core.
p3dr0 Posted April 6, 2011 Posted April 6, 2011 (edited) I think you have to add those ports (you connect your machine to) to one of your vlans to get it to work. Default vlan id in hp switches is vlan id 1 and you assigned 1 to your server vlan. Im not sure about that but its worth trying i forgot to add that your password is visible in cfg files you have posted i recommend changing and encrypting them Edited April 6, 2011 by p3dr0
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