Stack member switch monitoring

Unanswered Question
Apr 25th, 2007

Hi All,

Is there a way to monitor individual switch member in a stack? I mean like putting ip addres and using an icmp monitoring tool to monitor them 24x7?

What is the message in the log (and level) if one of the member switch fail? Will the same message retain in the log if the master fail? I'm thinking of configuring the switch to send its log to a syslog server, then run a log monitoring tool to monitor for that specific line in the log and trigger and alert via smtp, sms, or paging if icmp monitoring of member switch in the stack is not possible.


  • 1
  • 2
  • 3
  • 4
  • 5
Overall Rating: 4 (1 ratings)
sachinraja Thu, 04/26/2007 - 05:40

Hello dannilo,

As you must be knowing already, with 3750, you have only one management IP address on the master, which is used for monitoring all the other switches... this makes it tough to monitor the stack members with IP addresses.. if u want, u can configure one port on each stack member as a L3 routed port, instead of L2 and connect this port to the network management LAN, to monitor the switch, but this becomes too complex.. The best way to monitor the failure of a stack member is to use SNMP and send the traps to the network management station...

or else you can break the stack and cascade the switches to monitor all the sitches with IP, but this really negates the real use of 3750 stacking switches..

Hope this helps.. all the best.. rate replies if found useful :)


Danilo Dy Thu, 04/26/2007 - 06:36

Hi Raj,

Thank you for your reply.

Configuring one port of each member switch to L3 and connect to management switch is a good idea. However, it becomes dependent on management switch, besides, if the stack switch is in different segment and many hops away from the management switch where the icmp monitoring tool is connected its too complex and may become a security issue.

I have an idea which I may need help and advice if 3750 supports this. I'm thinking configuring one port of each member switch to L3 and connect a loopback plug. Will this works? Thus 3750 support hard loopback plug? Any command I need to execute in the interface?

If my idea works, the stack switches can be anywhere as long as reachable in the network by icmp monitoring tool, I would be able to monitor each member switch 24x7 :)

I think the designer/developer of stackwise forgot this basic but also important part of network - monitoring individual switch. I hope they will do something about this in the future code.


sachinraja Thu, 04/26/2007 - 06:44

Hey dandy

what exactly are u going to achieve by doing a hard loopback ?? anyway, u just need an ip address on the switch for icmp polling.. are u referring to loopback IPs here ?? I'm not getting it... yep, having L3 interfaces is complex, but just thought of telling it as an option.. best thing will be to configure a station which can receive snmp traps, so tht any switch going dowon on the stack will be notified soon...

Hope this helps.. all the best.. rate replies if found useful..


Danilo Dy Thu, 04/26/2007 - 06:51

Hi Raj,

Plugging an RJ45 loopback cable in the L3 interface to bring it up/up without plugging it on any host so I would be able to ping the IP Address configured in the interface :) like this one

Software Loopback interface will always be up as long as there is one member switch in the stack (that is a master), so it's not an option

I'll try snmp though and let you know. I prefer icmp because I use WhatsUp which never fails me yet for 7 years :)


sachinraja Thu, 04/26/2007 - 06:55

Hey dandy

yep u can try the loopback stuff.. hope it works for you :) try testing on a lab and see how it behaves first... let us know your result... close the post once u get all ur answers :), which can be helpful to others when searching


Danilo Dy Thu, 04/26/2007 - 06:56

Hi Raj,

Yup, I'll do that. Hope it won't screw up the interface :).



Danilo Dy Thu, 04/26/2007 - 18:48

Hi Raj,

Was working for a few seconds then the interface will go down to err-disable state. I don't find any workaround to prevent it to go to err-disable state and I found this;

"When a loopback cable is connected to a switch PoE port, the show interface status privileged EXEC command shows not connected, and the link remains down. When the same loopback cable is connected to a non-PoE port, the link becomes active and then transitions to the error-disabled state when the keepalive feature is enabled. There is no workaround. (CSCsd60647)"

From this link

I'm back to zilch now :( on how should I be notified if any of the member switch fails


This Discussion