1 Vlan out of 8 having issues

[mrbios]

Seem to be having some very strange problems with one vlan, up until 2-3 days ago it was working fine with no issues, i'm the only one that would/should be changing anything that could cause any problems () and i know i've changed nothing that could be causing this problem...

Started off with a complaint that the printer in the maths department was taking about 5 minutes to print anything, pinging the printer showed constant time outs then about 10-20 ping responses and repeat! First off it sounded like an ip conflict, so i tried changing its ip to another free one and that made it better but not perfect (which is very strange) then yesterday i had a similar problem with another printer (not connected to the same switch but in the same vlan) i frigged around with that for a bit and couldn't see what the problem was, then the power got switched off in that room (science lab) and when it came back one both printers were fine, constant ping responses .... ok i thought, must have been those two printers playing silly buggers with each other.

That was yesterday, not today i have an impero group with 44 PCs in, again on the same vlan as those 2 printers, and they're randomly going off and back on again on impero, but if i remote to one that goes off with vnc i can see the pc is working fine, i did test another machine with a constant ping and found that it did eventually time out and come back again - these unlike the printers are not on static IPs.

Some of the errors the DC for that vlan is picking up in the system event log:

Name resolution for the name domain.local timed out after none of the configured DNS servers responded.

obviously i've changed the real name of the domain there though for the sake of posting here. The DC is setup to point at itself and one other DC as secondary for DNS, wouldn't be worried about this if i had just restarted the server but this error was popped up during the middle of the day!

While transmitting or receiving data, the server encountered a network error. Occassional errors are expected, but large amounts of these indicate a possible error in your network configuration.  The error status code is contained within the returned data (formatted as Words) and may point you towards the problem.

Whether any of those are related to the problem im getting i don't know, the DC is running on vmware ESXi 4.1 in a cluster and is Server 2008R2, i've checked all the switches with procurve manager and they all appear fine too.

EDIT: did have 3 errors posted but removed the last one, fixed that.

EDIT: in closer inspection neither of those above errors seem to be the cause, i might just try restarting switches now to see if they're causing a problem anywhere, nothing in any of the event logs though ¬_¬

[Geoff]

Can you give us more info on how your vlans are setup, what switches are involved and how they are all connected together?

[mrbios]

The switches are HP Procurves, core is an 8206zl, fibre connection from that to a 6108 which is the link into that block then there are 4 HP 2650/2610s which span off that 6108 all via copper. The vlan in question is untagged throughout with a print vlan, tv system and door system vlans tagged through to appropriate locations. (should note the 2 printers i mentioned aren't on the print vlan, but there is one which is and that hasn't had an issue.....as far as i know)

I'm thinking someones connected a loop somewhere but surely the switches would pick that up?

[Geoff]

Only if you enabled STP and loop detection (which you should!).

http://h40060.www4.hp.com/procurve/u..._Dec_08_A4.pdf

To narrow down the issue I would suggest you assign the switches involved an IP on the problem vlan and try and ping them over that vlan. When you start getting dropped packets you know you've just gone over the problem network link. This will at least allow you to narrow it down to a couple of switches and their associated cabling.

[mrbios]

I've enabled STP on the switches that weren't enabled, two switches keep blocking port 50 which is the link back to the 6108 for access to the rest of the network.....the only reason i could ever see them to do this is if they had a connection to each other and they were incorrectly reading the best route back but they're so far apart and there's no links between them

I have noticed i have an 1810-24G i can't access over there though so I'm going to go and poke that one a bit.

[Geoff]

check the mac address tables on the problem switches for the other switches mac address range. Then you can work out whats connected where. Alternatively you can enable LLDP and use that to map the links.

If you check the switches system logs you can find out when and why it disabled a port.