I'm stumped

[Lee_K_81]

Ok, here's the gist of it:

 

We have a network that was happily chugging along, 160 thin client introduced in summer, working ok.

 

Now:

 

all of a sudden everything is screwed. No changes to any config anywhere. CMIS keeps booting people out, Thin Clients freeze and loose connection to TS. Checked error logs on TCs message saying "The Group Policy client-side extension Folder Redirection failed to execute. Please look for any errors reported earlier by that extension." but that seems to be logged before the thin clients log on.

 

My thoughts, and things i'm currently looking at are:

 

Dodgy nic on DNS server - could loose connection when Thin Clients are trying to connect? Nothing in DNS logs.

 

Failing Switch - seems most likely as it's quite old (Cisco Catalyst 3500) and it is the switch that most servers are plugged into either directly or indirectly.

 

We are actively monitoring the servers, and managed switches including the perhaps dodgy cisco with nagios. Other than the occasional warning bcs no data recieved from host everything seems to be fine. We no data recieved warnings disappear on next polling.

 

Does my hypothesis seem correct? It's very frustrating as all the equipment is telling me it's fine, but something is going wrong somewhere. We've managed to take the cisco out of commission so hopefully tomorrow will tell. If not suggestions of other things i may be over looking would be appreciated!

[localzuk]

How about a network loop? These cause weird and not so wonderful things to happen?

[tech_guy]

Yes, do you have hubs/switches in public areas. The little darlings might have been fiddling?

[Geoff]

You may also have a cable fault. Basically someone being careless with an electric drill.

[powdarrmonkey]

One of my GPOs got corrupted once and started playing havoc with absolutely *everything*. Some really weird stuff that you'd never think was related.

[Lee_K_81]

How about a network loop? These cause weird and not so wonderful things to happen?

 

We did think that at first but the behaviour isnt the same.

 

You may also have a cable fault. Basically someone being careless with an electric drill.

[/Quote]

 

As far as i'm aware no body has been drilling anywhere on site recently.

 

Yes, do you have hubs/switches in public areas. The little darlings might have been fiddling?

Only one behind reception. Dont think they'd have managed to get behind there. Only people who go in there are the receptionists and SLT.......... Maybe i'd better double check that cab.

 

One of my GPOs got corrupted once and started playing havoc with absolutely *everything*. Some really weird stuff that you'd never think was related.

 

That one sounds interesting, i'll keep an eye on that one, did anything show up in event viewer anywhere to give any clues?

[Netman]

Have you tried dcdiag and netdiag? They're pretty good at picking up on any problems...

[m1ddy]

try running a network protocol analyizer. If there are a lot of broadcasts coming from the same IP/MAC or broadcasts coming from everyplace then it could be a faulty NIC killing your bandwidth or a loop somewhere in the topology.

 

Are you running spanning tree?

[powdarrmonkey]

It did log some errors, but I don't remember what. It was quite a while ago.

 

They were pretty misleading though, we spent about a fortnight chasing around trying to work out what was wrong with the damn thing.

[PEO]

You may also have a cable fault. Basically someone being careless with an electric drill.

 

ha yeh i've had that happen to me lol

[Lee_K_81]

Thanks for all the suggestions, had tried running wireshark, implemented spanning tree, all yielded no results. Managed to swap the suspected faulty cisco switch out (which is al LOT harder than it sounds when you have no spare switches), asked staff not to use thin clients in 3 of the suites so we could test in a smaller scale, all seemed well. Alll seems to be running well right now! Phew!

[Matstones]

Ok, here's the gist of it:

 

We have a network that was happily chugging along, 160 thin client introduced in summer, working ok.

 

Now:

 

all of a sudden everything is screwed. No changes to any config anywhere. CMIS keeps booting people out, Thin Clients freeze and loose connection to TS. Checked error logs on TCs message saying "The Group Policy client-side extension Folder Redirection failed to execute. Please look for any errors reported earlier by that extension." but that seems to be logged before the thin clients log on.

 

My thoughts, and things i'm currently looking at are:

 

Dodgy nic on DNS server - could loose connection when Thin Clients are trying to connect? Nothing in DNS logs.

 

Failing Switch - seems most likely as it's quite old (Cisco Catalyst 3500) and it is the switch that most servers are plugged into either directly or indirectly.

 

We are actively monitoring the servers, and managed switches including the perhaps dodgy cisco with nagios. Other than the occasional warning bcs no data recieved from host everything seems to be fine. We no data recieved warnings disappear on next polling.

 

Does my hypothesis seem correct? It's very frustrating as all the equipment is telling me it's fine, but something is going wrong somewhere. We've managed to take the cisco out of commission so hopefully tomorrow will tell. If not suggestions of other things i may be over looking would be appreciated!

Have you upgraded your terminal servers... If you have Linux based thin clients the open source RDP client in your thin-clients may no longer be 100% compatible with the servers