Capture/log port issues on switches

Unanswered Question
May 23rd, 2010
User Badges:

Hello,


I have a couple of 3750G's that I'd like to point to my syslog and log most issues above the debug level, how can I do this?  I think I have configure the switches to point to the syslog server but I need to set the logging level and I guess the buffer size?


Also what do engineers use to be alerted or be able to centrally montor collisions, crc, runts and STP issues on switches?


Many thanks

  • 1
  • 2
  • 3
  • 4
  • 5
Overall Rating: 5 (1 ratings)
Loading.
hobbe Sun, 05/23/2010 - 13:49
User Badges:
  • Gold, 750 points or more

Actually I think you are  trying to do two different things with your switch,


Monitor and log.


To monitor a switch you measure the different variables that you want to know something about.


To log (via syslog) you only get a syslog message telling you that X have happened.


To monitor you can use SNMP and both send traps or poll the device.

This is very good for fx checking how much data is transfered over an interface and get graphs over time of that information

or how much CPU is working, or memory, or temperature and such information.

basically things you want to have graphs of over time.


Logging is a different matter alltogeather.

logging is very good for things like checking if an access-list is hit, if there is an alarm going of for temperature or if the switch restarts, or if someone changes the configuration and such things.

In general things that does not happen that often and you do not need to get graphs over time for it.


To monitor I use SNMP to a monitor software that will show me the results in graphs and numbers.

if there is a problem the software will set of an alarm and send me a SMS and/or e-mail that something is not right.


To log I use a Syslog server that will let me log everything that I want and lets me filter the logs and show me what I want to see.


My general rule is to log everything and then filter out what you need for the moment.

if there is need for more logs afterwards then it is there.


Good luck


HTH

Andy White Sun, 05/23/2010 - 14:05
User Badges:

Ok thanks, to make it simple say I just want to get some sort of alert should a port have a runt, collision, etc. I currently have snmp running that is doing the cpu, mem etc to my orion solarwinds npm server.

hobbe Sun, 05/23/2010 - 14:24
User Badges:
  • Gold, 750 points or more

Then I (I think) you already have it.


The NPM is the Solarwinds Orion NPM right ?

To get an alarm out of solarwinds NPM you will need to configure the alert manager.

it is a separate program on the Orion NPM server.

But the best view I think you should have by just checking the top 10 page and if 10 is not enough then you can turn it up to fx top 100 and so on.

easy to check for whoever is in charge att the moment.


HTH

Actions

This Discussion