Printing/Network Issue site wide - Help!

[getoffmymac]

Hello,

 

Long term lurker first time poster

 

I work as a technician in a fairly large secondary, my network manager is sadly on long term sick

 

Recently I've been having an issue where all of my printers on my print server (server 2008R2) go offline, the devices are not pingable from the print server

 

The printers are on their own VLAN and from what I can tell nothing outside of this is affected.

 

There is to my knowledge no configuration changes that have taken place, I've done a walkaround the best I can and cannot see any physical loops however.....

 

In desperation I powered down all of the switch stacks today across the site and printing seemed to return to normal, I could print, ping the devices from the print server this lasted for an hour or two however this evening I've dialled in to check and it appears the issue persists. I can't ping the devices from the print server or/and at the very best i might get a response every once in a while.

 

My next plan is to run continious pings and power down the cabs one by one, wait - and see if this changes anything. Is there anything else I should be looking for??

 

Please be gentle as some of this is fairly new to me (server management, vlans etc)

 

Many Thanks

 

Shaun

[Mr.Ben]

Are you using HP switches? You can configure a loop protection feature to find the culprit.

[kmount]

Can you ping the gateway of the printing vlan when it goes wrong? I'm wondering if something is stealing the gateway IP and the whole lot dies because of this.

 

Also, put a pc onto that vlan and run pings from there too when it goes wonky so you can see if traffic still works within the vlan (that will help give you geographical representation too)

 

The logs on your switches will be key too to see if there's something happening with spanning tree or whatnot.

[Davit2005]

Spanning tree could be a cause.

 

Are there any logs on the core switch where the vlans are routed?

 

You could run a packet capture on the vlan to see if you can spot anything.

Edited November 22, 2019 by Davit2005

[MrWu]

Failing that do it the old fashion way ... depending on how many printers you got, unplug them one by one and narrow it down to see if it a printer causing it, and check if any other devices are plugged into that VLAN

[getoffmymac]

Thank you for the replies guys - its been a long day but I was scared to have this hanging over me on monday

 

As it stands now I (think) I've narrowed it down to a single cabinet/switch stack.

 

I restarted this stack and now can't get on it via IP (they are dell switches) but can access via console, since this stack when down and hasn't come back up properly it would appear printing is stable - hopefully not concidence.

 

I can now ping addresses in the printing vlan with no latency so my thoughts are its either the fibre to the stack, the stack itself, a dodgy spf module or the like.

 

I tried unsuccessfully today to power on each switch one by one but there is no connectity (cant ping the stack and stack via console can't connect to core)

 

Cores seem to be happy and all other edge stacks are accessible

 

Wish me luck - any further ideas apppreciated

 

Thus far

 

Tried different gbics

Tried different fibre cables

Tried powering on switch one by one

[getoffmymac]

Ok update time again

 

Think Ive narrowed it down to one of two things

 

The stack config is just buggered

Spanning Tree or the like is blocking the stack from communicating?

 

I swapped over fibres and various other bits today but to no avail

[kmount]

Can you get the config from it?

[getoffmymac]

Can you get the config from it?

 

Yes I believe I can, Ive just been comparing them via console connection/putty - Ill dump it out tomorrow as I powered the stack offline before heading home this afternoon

[kmount]

I'd be looking in the logs on both the stack and the core to see about spanning tree messages.

 

I'd also check generally the configuration in terms of bridge priority but also looking at if you give the stack a different IP can it reach the network then?

 

It would be worth simplifying things considerably too by unplugging the access ports on the stack if you can so you're left with empty stack connected to network. Does the problem still occur?

 

Then gently progress adding ports into the mix whilst watching logs and experience.

 

Hopefully you're going to find the smoking gun.