IT Question: LAN Monitoring

I totally agree. You can have hierarchical layer two networks remain stable beyond 2000 hosts with proper use of storm-control. It only takes one switch to have STP misconfigured on a small network and you end up with a switching loop. But as you say it would probably bring the network down rather than slow it... but this is not always the case, some pre-configured switches use storm-control which can help control broadcast storms but this would still lead to congestion and reduction in network performance.

Originally Posted by grimey121uk

If he is using managed switch he could have snmp up and running in 30 mins with stats on every single port for traffic in/out as well as speed/duplex issues, plus you can break traffic in/out into broadcast, multicast and unicast as well as showing the number of frames dropped

Do this. Have a look at http://www.paessler.com/ They make a tool called PRTG which I have used in the past, quite useful.

There no right or wrong way to diagnose it,

Me personally (if the switches are snmp capable), I would get a windows server download PRTG network monitor on a 30day trial, it takes 15 mins to set up initially enable a community string on the switches punch the management IP into PRTG which can autoscan the device so if it is a "common brand" switch it will automatically monitor all switch ports with a few clicks giving a full picture of the entire switched network.

I work in a network operation center and some of our customers networks contain 100s of switches and without snmp they would be totally unmanageable

Originally Posted by ravingdavis

http://www.paessler.com/ They make a tool called PRTG which I have used in the past, quite useful.

I use PRTG at home as my home network is quite large so I use snmp to monitor it if the internet goes down I can jump on my network maps and instantly see what device is down etc - http://www.ants-network.co.uk

We use nagios at work but its a bit tricky to configure if your not a linux buff

If you just want a quick overview then setup a mirror port on the uplinks and use ntop.

It will give you pretty graphs and protocol breakdowns without the ball ache and fine tuning required to setup netflow/cflow.

You can also use cricket/cacti to monitor the throughput of each port using SNMP, but it won't give protocol breakdowns.

I doubt it's a L2 loop, as others have said it would be a notwork not a network. Hint: If *ALL* the activity lights for a vlan go from happily blinking to solid then you have a nice L2 loop/broadcast storm. Oh, they are fun, really.

Wireshark is great if diagnosis but I wouldn't use it for an overview of what is going on.

Dan

Took you two pages of debate to get to PRTG? You bunch of amateurs! Thought you guys were network gurus?

thanks for the response guys but wireshark still looks like a dark art to interogate, and have also been recommended Solarwinds netflow & performance monitor, which aparently is a little more acceptable to use if your not a network engineer.

grimey121uk: I have a layer 3 HP procurve switch 5412zl, serving all hosts on the network, which is around 85 pcs....10 printers....and unfortunately a multitude of iphones using Wi-Fi facilities.

I am looking at ways to segement the Wi-Fi traffic onto a separate VLAN in order to control what the Wi-Fi users are surfing, and this too could be hindering my network.

Originally Posted by -Ralph-

Took you two pages of debate to get to PRTG? You bunch of amateurs! Thought you guys were network gurus?

Ralph: is the PRTG getting started guides fairly straight forward to set-up?

Originally Posted by slloyd

Ralph: is the PRTG getting started guides fairly straight forward to set-up?

its dead easy mate, you just need to set your switches up with an snmp community string

Any of the above tools are easy enough if you are prepared to take your time and RTFM. If you are a server/desktop guy then networks can seem like a bit of a black art, but I was once in that position and nobody taught me. Don't expect to be able to plug and play a tool and it present you with the answer. Getting data is quite easy, understanding what it means is something else.

I'm a big fan of RTFM. You will save time in the long run if you sit down and RTFM, and when you come across bits you don't understand google, or get yourself something like a networks for dummies book and sit and read it. How much time have you spent trying to fix this already? This thread has been running for 4 days. During which time your office is having problems, the bosses are ****ed off and you look sh!t because you are not fixing it. It may take you a day to lock yourself away somewhere, but that's better than a day farting around blind downloading and installing tools and fixing nothing. You'll be able to tackle stuff like PRTG with confidence, and the next time you have a networking issue you'll be much better placed to tackle it and fix it straight away.

Oops, only just spotted this.

+1 to PRTG, tis superb.

Originally Posted by slloyd

thanks for the response guys but wireshark still looks like a dark art to interogate, and have also been recommended Solarwinds netflow & performance monitor, which aparently is a little more acceptable to use if your not a network engineer.

grimey121uk: I have a layer 3 HP procurve switch 5412zl, serving all hosts on the network, which is around 85 pcs....10 printers....and unfortunately a multitude of iphones using Wi-Fi facilities.

I am looking at ways to segement the Wi-Fi traffic onto a separate VLAN in order to control what the Wi-Fi users are surfing, and this too could be hindering my network.

Netflow is mainly cisco, I would imagine HP use slflow, don't forget flow technology will only recognise traffic when it crosses a layer 3 boundary which won't give you a full picture of any issues on a L2 switched network.

PRTG also supports s/net flow, it will take you less than an hour to configure