11 Replies Latest reply on Jan 24, 2014 5:31 AM by gbggri1

    Cisco 3745 and 3845 router monitoring power supplies

    dsteuerman

      A bit of a newbie but managed to become fairly familiar with the alert config thus far.

      I have read through most of the posts with respect to Cisco chassis based switches and how to monitor/alert where multiple PS are present. I have been doing a bit of tweaking myself and after a large number of combinations, managed to get the results I wanted (but haven't fully tested). For all switches with single PS, I get the email and event for current switch "Operational Status". For chassis based switches I get the same for both PS (I.E from an alert email ***Node XXXX IP address x.x.x.x Power supply state has changed to status UP, down***) same for all my 7200 routers with dual PS, 4500's, 3750 etc...

      The problem:

      I have multiple WAN router platforms; some with single PS and some with dual. The ones with dual not only have dual PS but AUX -48V DC PS ports if you were to use a redundant DC PS, which I'm not. The problem is those PS are listed when the Poller polls these routers and produces a 1,1,5,5 for their status. 1,1 for the AC PS and 5,5 for the non existing DC PS.

      I've tried multiple rules, combinations to try to ignore this power supplies status, I can’t find anything with regards to admin down these unused DC PS ports etc...

      I'm using the ciscoEnvMonSupplyStateTable w RAW VALUE and for the format I'm using the Operational Status (operational status gave me the word UP/DOWN/UNKNOWN opposed to the numbers).

      My alerting is very simple:

      Poller name_is equal to_ciscoEnvMonSupplyStateTable

      Status_is not equal to _UP...

      So when I applied this nothing alerts except all my 3745's and 3845's....Any ideas??????

      Output from one of my 3845's.

       SYS PS1 is present.
              Fan status: Normal
              Input Voltage status: Normal
              DC Output Voltage status: Normal
              Type: AC
              Thermal status: Normal

       SYS PS2 is present.
              Fan status: Normal
              Input Voltage status: Normal
              DC Output Voltage status: Normal
              Type: AC
              Thermal status: Normal

       AUX(-48V) PS1 is absent.

       AUX(-48V) PS2 is absent.

      The email I get for the PS status is:

      ***Node XXXX IP address x.x.x.x Power supply state has changed to status UP, UP, DORMANT, DORMANT

       

       

        • Re: Cisco 3745 and 3845 router monitoring power supplies
          Steven Klassen

          It'd be nice to be able to do this with one poller & one alert, but it looks like you're not going to be able to. What about having a separate alert for your single PS and another for your dual PS with a custom node property to tell one from the other?

          1 of 1 people found this helpful
            • Re: Cisco 3745 and 3845 router monitoring power supplies
              dsteuerman

              The problem isn't with one PS devices vs 2 PS devices. The way I have the poller and alerting setup, everything works fine for say 3750E switches, 6500 switches w/ dual PS etc...The problem is with the 3745 and 3845 specifically where the AUX -48v DC power supply PORTS that have nothing connected to them are reporting the port status as DORMANT. So basically when the poller polls these devices it polls 4 PS and reports 2 as UP and 2 as DORMANT.

              1 of 1 people found this helpful
                • Re: Cisco 3745 and 3845 router monitoring power supplies
                  pyro13g

                  I'd bet you are going to get the same type of message when one of the PS in your other gear goes to any other status than normal.

                  Just take advantage of Syslog or Traps. 

                  What really needs to happen is the table needs to be walked to find out how many PS entries there are and report the status of each row individually.  Or, you need to make a specific poller for each row and create alerts based on each row.

                  We just use Syslog, works fine.

                   

                   

                   


                  1 of 1 people found this helpful