Run report to find failed POE switches?

  • 6 July 2017
  • 6 replies
  • 473 views

Userlevel 2
Hello -We have x460-48p gen1 switches system wide. When one fails POE we know about it b/c poe devices like cameras or WAPs are down which we monitor. However there are some switches whose POE fails that we don't have any POE devices on so it's not as critical. But we want to replace these for when we do need them in the future to provide POE. As we notice them we replace them, which doesn't seem to be a most efficient approach. Without logging into each stack and doing the traditional show inline power and deducing from that. Is there any way I could run a report using Netsight or similar to query all the switch stacks and report back slots where POE failure is seen?

Some things to consider?

Here's an example of a stack of 5.

# sh stacking

*1 Active Master CA-

2 Active Backup CA-

3 Active Standby CA-

4 Active Standby CA-

5 Active Standby CA-

when you do show-inline power and you have a failed POE device, the full list won't populate.

In this example slot 3 above has failed POE so only slots (1 in this case is a non POE switch) with poe UP to the failed slot will show, not beyond.

# sh inline-power

Inline Power System Information
Configured : Enabled
Power Usage Threshold : 70 percent (per slot)
Disconnect Precedence : deny-port
Budgeted Measured
Slot Inline-Power Firmware Status Power (Watts) Power (Watts)
2 Enabled Operational 760 W 0 W

See how it only shows inline output up to the switch that is working properly. You know that slot 3 is the one that's failed. But more than that could have failed, so to confirm I'll do a show inline-power info ports 4:1-10, 5:1-10 to make sure that they are searching for power.

Without having to confirm each site and the switch stacks at each site for how many are in the stack, and then querying show inline-power command and cross referencing for what's showing up is there a way I could somehow run a report for which slots in a stack have non-working POE?

Sometimes in the logs I'll see poePlusInit: PSE initialization failed other times bcmPseReceiveMessageString: timed out

Thanks,

Sarah

6 replies

Hi I faced the same issue If you run diaganstoic extend, then the switch will do extend diagnostic and reboot. Then #sh Diag will show POE module failed. Then you could know that switch PoE is failed. Thanks Khaled
Userlevel 2
Thanks Khaled - we are familiar with that process and do use it. I am more looking for a non time consuming non reboot required method to collect data from current info? Like maybe leveraging netsight? To run extended diag on all slots on all stacks across the enterprise would be quite time consuming 🙂 I'm trying to find a way to gleen from information already on the slot and focus on that. Kind of like high level get the information on where the issues are seen to be able to directly address. Ever had luck with something like that?
Userlevel 3
We can configure/run a script in switch which match the log error message. But you need these information to be viewed on NetSight for all slot ports?
Userlevel 2
We can configure/run a script in switch which match the log error message. But you need these information to be viewed on NetSight for all slot ports?
I think so, b/c it could be on any slot in the stack in theory?
Userlevel 7
I was also thinking of a script, preferably python (15.6+). Something like: Check which slots are PoE Do a sh inline-power per slot Check result With exos 21.1+ (sorry, I know you can't here), I'd jsonrpc it to query any stack from a server/host. You should be able to manage the same using cli scripting and ExtremeManagement.
Userlevel 7
I was also thinking of a script, preferably python (15.6+). Something like: Check which slots are PoE Do a sh inline-power per slot Check result With exos 21.1+ (sorry, I know you can't here), I'd jsonrpc it to query any stack from a server/host. You should be able to manage the same using cli scripting and ExtremeManagement.This is the way I would do it. It would be much easier with Python, but should still be doable via TCL on EXOS < 15.6.

Reply