Symmetra RM 6000 -- Am I on bypass or not?

[Anonymous user]

I have a (very) remote Symmetra RM 6000 with a single battery installed (don't laugh, it's what the client provisioned.)

The web card was reporting two issues: first it was reporting a faulty battery and second it was (and still is) reporting a system level fan fault. It was also reporting that it was on bypass, ostensibly due to the faulty battery.

Today I got someone to replace the battery, so that problem is gone. The System level fan fault still exists.

After we replaced the battery and took the UPS out of maintenance bypass, there was a system event saying "UPS: No longer in bypass." In the web interface, there is nothing to indicate that the UPS is in bypass and on the Battery System Status page it shows 19 minutes of runtime remaining (I said don't laugh!!) So, I am lead to believe that the UPS is no longer in bypass.

The confusing part is that we have a Nagios SNMP check polling upsBasicStateOutputState which is saying that the UPS is in bypass. According to the MIB, this OID is "an ASCII string containing the 64 flags representing the current state(s) of the UPS." Flag 16 is returning 1. According to the MIB, this means that the UPS is "In Bypass due to Fan Failure."

So which is it? Is the UPS actually in bypass and just not telling me via the web interface, or is the MIB's documentation incorrect?

Thanks for any clarification!

[BillP]

Hi Sean,

Can you tell me what firmware you have on the Network Management Card (NMC) in the UPS? That way I can know what you're working with. I'd also be interested in any log files you can obtain from it (which will depend on NMC model and firmware version). Instructions here on downloading logs -> http://www.apc.com/support/index?page=content&country=ITB〈=en&locale=en_US&id=FA156131 .. you can attach them to this post if you're able to obtain them successfully.

Did any on-site person happen to note which LED(s) were illuminated on the UPS display?

[Terry_Kennedy_apc]

By the way, the fan is an orderable spare, part number 490-0024B. You can get it from many of the usual places (CDW, etc.) although it is usually not stocked and they have to order it for you. The recommended replacement procedure is to shut down and de-energize the UPS (removing the fan leaves a big opening in the UPS where dangerous voltages are present, and you have to reach in there to get to the fan's plug). It is possible to do it on a live UPS (at your own risk) if you put the UPS into bypass. While nothing should happen if you do it without being in bypass, any glitch can "scare" the Intelligence Module and trigger a fault, which may drop the load.

[Anonymous user]

Hi Angela,

The NMC is an AP9631 running AOS v6.2.1. I have attached the (sanitized) event log and and the data log. The fan fault happens at 01/18/2016

Unfortunately our on-site person did not note which LEDs were illuminated.

[Anonymous user]

Thanks, Angela! 

On 1/20/2016 10:53 AM, Angela said:

I don't see one of those 64 flags that just says "fan fault" so I am guessing maybe it shows this one since its the only option to represent the fault.

That's what I was thinking. Hopefully one of your Symmetra RM folks can confirm for us!

The output of upsBasicOutputStatus is 2, which is 'onLine' so that is encouraging.

I have attached screenshots of the Home page and the Control -> UPS page.

Thanks again for your help!

[BillP]

Hi Sean,

I confirmed Symmetra RM does not go to bypass when there is a fan fault. Based on all of this, I think flag 16 is just a little misleading and the UPS is online.

As a user with a System Fan Fault, would you prefer to see no flags triggered on the upsBasicOutputState or now that you know, continue to see the flag 16 which is a little misleading since it says in bypass due to fan fault? I think ideally, the answer is to add a new, clarified flag that just says system fan fault to the OID responses but I was just thinking if I ask someone to make adjustments (if it is even possible), what is preferred by a user and what you'd expect ideally as well as with what we have available now to us?

My gut tells me flag 16 is better than no flag at all, especially for those who only monitor via SNMP polling.

[Anonymous user]

Angela,

Agreed, flag 16 is better than no flag at all for our purposes. We can adjust our Nagios check to work with this new information. I also agree that a new, clarified flag would be the ideal solution but in lieu of that, some easily searched documentation in the knowledge base or an amendment to the user manual, or even a note in the MIB itself -- space permitting -- would also be fine.

Of course, asking on the forum evidently works quite well, too 🙂

Have a great day and thanks for all your help!

[BillP · Accepted solution]

Thanks for your feedback. And you're welcome!