Questions about 3com network supervisor.

[VitaNatura]

Hi everybody,

I'm using 3com's tool to minitor our network. A lot of the following events are logged:

Severity Time Resolved Name Type Description Comment Explanation

Info 12/12/2006 13:25 Not applicable 3Com 4200 48ports SE (r) 3Com SuperStack 3 Switch 4226T Trap: Link is available on unit 1, port 16 A link on 3Com 4200 48ports SE (r) is available. The link is connected to "dtp0410-125" on unit 1, port 16.
Info 12/12/2006 13:24 12/12/2006 13:25 3Com 4200 48ports SE (r) 3Com SuperStack 3 Switch 4226T Trap: Link is unavailable on unit 1, port 16 "A link on 3Com 4200 48ports SE (r) is unavailable. The link is connected to ""dtp0410-125"" on unit 1, port 16.
Check the cable is securely plugged into both ends.
Check the cable is of the correct type (cross-over or straight through), length and category (e.g. CAT 5), and that the cable is not faulty.
Check that the Ethernet adapters at both ends of the link are working.

This event has been logged with a reduced severity of ""Info"" because the associated link is not currently being monitored."

I would like to know if this indicates that the link actually is going down? I think it's weird, because it says it's unavailable and immediately after that it's available again. The above is just an example, he logs this message for the same switch, but different ports, one after the other, however it doesn't do it for all of the switches ports.

Thanks in advance

 

[SYQUEST]

It is a possiblity you have some flaky patch cords or cabling or bad jack to a workstation or other device. An intermittent could cause that type of response. I would checkout the items mentioned above for defectives. In fact if the patch cords were manufactured before 2003, I would replace them with a new set.

The reason being in 2003 test standards were put in place for patch cord manufacturers to meet performance specs of the standard on a verifiable basis with some new test and manufacturing procedures.

Hope this helps!

....JIM....

 

[RobKorn]

These messages from 3ND are info. only, no need to research them unless you are having "reported/observed" problems with the connection. You can generate this message by disabling the NIC in a workstation, rebooting the workstation, or any other event that closes the port.

I generally disable that message (both the unavailable and the companion available) to limit clutter in my alerts area. As it is, I get 4000+ alerts a day... hard to see the real problems when you are bombarded with "info".

Good luck.

 

[VitaNatura]

@RobKorn

I assume that you frequently use 3com network director?

Quiet often it happens that there is a cloud with a lot of devices connected to that cloud. While some devices are connected to our central stack (consisting of 4 3300 series switches) and a 3848 switch, the map is not accurate.

Do you also have some tips about how I can get a clean map when running a new scan?

Thanks in advance

 

[RobKorn]

The clouds you speak of are a pain. Sometimes they are caused by having "non managed" devices on the LAN. Also, you may have "managed" devices that have incorrect SNMP community strings.

One of the problems I have is with "mobile" devices. If I scan today and a client machine is found on "AP1" then scan later and find same client on "AP2" the software makes a mess of it. Even goes as far as to report what looks like dual connections to the client. I don't know the solution but there is a setting in Options/Discovery that refers to "preserve topology" that may be a key. Haven't tried playing with it though.

Because of the size of my network, I couldn't do a complete (all segments/all addresses) scan. Instead I maintain a text file of my segments with specific ranges of addresses to scan. This limits the clutter. Let me know if you need help on setting that up.

Good luck.