Wired Intelligent Edge

Has anyone been experiencing POE power supply failures in the MAS 2500 product line? We have been getting them RMA'd but are noticing it happening more frequently.

Power Supply                  : Failed (580W)
                              : 12V System Voltage Ok
                              : 56V PoE Voltage Failed

Thanks,

If you are seeing it frequently particularly in the same locations it may not actually be a failure of the 56V rail but the protection circuit. There is a protection circuit in the power supply that if the unit encounters an anomaly in the 56V rail like over current, over voltage, etc, then the 56V rail will shutdown but the 12V will stay up so that TAC/Engineering could possibly debug if needed. When in this condition, the unit must be power-cycled (e.g. removing the cable) for the protection circuit to disengage. A "reload" will not disengage the protection circuit. If the 56V rail does not recover after a power-cycle, then there is something else wrong with the switch and an RMA should be requested. Have you tried just removing the power feed and re-applying to see if it resets? If it does, I would look for other sources of issues such as possible bad weather at the time of the event, incorrect cabling, etc.

Thank you for your response. That was the issue. After pulling the plug and bringing back up the POE power was there.

I have another question related to this. We have about 400 switches along with Airwave. Is it possible to setup an alert to be nofified when the 56V POE fails?

I notice this in the switch logs:

Aug 19 04:07:26 :335304:  <WARN> |ChassisManager|  Major alarm: power supply failed on slot 0 at 2015-08-19 04:07:26 (cst)

Thank You

netbum

As far as I know Airwave doesnt have any triggers for syslog, but, you have a couple of options to automate it.

#1 write a script that interrogates the syslog database (aka device_event table) within Airwave, run it from the shell. For getting it working quickly, just test it as shown below. Of course running a string match on such a large table is not really a good idea (it would be good to modify the query further to add more where clauses to narrow it down, i.e. once you have a hit, note the severity and add that as a further "where" clause), but it should be ok for a once a day query.

Here are two different queries to try, the first is just for syslog (type=200), over last 24 hours (adjust as you see fit, i.e. change the '1 day'). dbc is a shell alias, you need to log into Airwave via ssh as root.

# dbc "select * from device_event where timestamp >= date_part('epoch'::text, now() - '1 day'::INTERVAL) and type=200 and message like '%power supply failed on slot%';"

 source_device_id | ap_id | client_mac | message | severity | facility | category | type | timestamp 
------------------+-------+------------+---------+----------+----------+----------+------+-----------
(0 rows)

#

and here is another query - I believe the MAS should send a power supply fail trap, you can try to find it using the following:

# dbc "select * from device_event where timestamp >= date_part('epoch'::text, now() - '1 day'::INTERVAL) and type=100 and message like '%PowerSupply%' ;"

 source_device_id | ap_id | client_mac | message | severity | facility | category | type | timestamp 
------------------+-------+------------+---------+----------+----------+----------+------+-----------
(0 rows)

#

#2 install a standalone syslog that does support message triggers, and/or cobble together something from a syslog + scripts that check for certain messages. Using something like graylog + rsyslogd would make this quite easy to do (graylog is quite a lot easier than logstash for handling arubas (err, malformed) syslog messages.

note - I should mention that the above is with a view to somehow automate checking+alerting, you can always manually search via webUI -> system -> syslog and traps -> message (box on right).

regards

-jeff