NMC2 does not detect UIO probes if UPS is not responding

[Terry_Kennedy_apc]

[Unrelated to my other question about Symmetra RM unexpected shutdowns, I just found it in my queue and remembered to post it.]

The AP9631 will not detect / report on any attached UIO probes until it can successfully communicate with the attached UPS (sumx or sy, depending on the firmware loaded on the NMC2).

This is a rather odd condition, so let me explain... I use either an AP9617 with an AP9312 module and AP9312TH sensors in a triple expansion chassis and power supply, or an AP9619 with an AP9312TH sensor in a single-slot chassis with power supply as a stand-alone environmental monitoring system. I had started with AP9319's, but those were orphaned a week after I bought them (the AP9319 is so orphaned that it doesn't even show up if you search the APC site for it). So I built my own, long before NetBotz. The only cosmetic issue is that the web page reports "Searching for a UPS", but SNMP monitoring / traps from the sensors works perfectly.

I decided to try using an AP9631 (firmware 6.2.1) with AP9335TH sensors, again in the single slot chassis (AP9600) with power adapter (AP9505). It does not detect the sensors. There is no listing for "Universal I/O" in the "Status" tab (nor is there one for "UPS", but that is expected).

I installed that same AP9631 and AP9335TH's in a Symmetra RM and the Universal I/O is detected, along with the UPS. I did not test to see what would happen if the NMC2 loses communication with the UPS - the Universal I/O may continue to work, I'm not sure. It appears that this is a regression from the older AP9617/8/9 cards, which did detect the sensors even when no UPS was present. 

The next time someone is working on the NMC2 firmware, would it be possible to check and see if this is easy to fix? I'd like to upgrade these to NMC2's for the improved security features, but I'm not willing to buy a UPS for each to make them happy  - these are at sites with multiple giant (500KW and larger) facility UPSs in place.

[BillP]

Hi Terry,

Take a look here -> http://www.apc.com/support/index?page=content&country=ITB〈=en&locale=en_US&id=FA156184 

You can probably get it working on v5.1.7 per that kbase but I did have a bug logged on this behavior for v6 firmware as it seems it never got ported over. I will see if they can squeeze it in for the next release since maybe they can port the fix from v5.1.7.

[BillP]

Hi Terry, just to confirm, we will have it addressed in the next revisions of sumx and sy applications as the developer was able to squeeze it in. Now I just need to make sure it makes it to all apps that support UIO. 

[Terry_Kennedy_apc]

That's wonderful, thanks for the lighning fast service!

[Anonymous user]

Hello  Angela 

I am experiencing the same issue here with an AP9631 running v6.5.6 AOS and sumx application v6.5.6. It is installed in an AP9600 chassis with no UPS connection. The NMC reports "There are no sensors connected" despite the presence of an AP9335T on the UIO port. Can you confirm that this issue was addressed and suggest next steps for troubleshooting?

[BillP]

Hi Patrick,

I just spent a little time looking at this. We did fix this but I see as of 6.5.6, it appears to be broken in some way again. As late as v6.5.0 it works.

So, the best option I see is to use AOS/sumx v6.5.0 if you need this to work immediately. In the meantime, I will see if we can fix it for a future release. We have a 6.6.4 coming in a few weeks but I think it will be too late in the development process to fix it in that rev unfortunately but I'll see - just wanted to set expectations.

If you want to load 6.5.0 sumx, you can get it here: ftp://restrict:Kop$74!@ftp.apcc.com/restricted/hardware/nmcard/firmware/sy/650/apc_hw05_aos650_sy650... 

[Anonymous user]

Thank you for the fast and comprehensive response, Angela. I'll give version 6.5.0 a try and see how that goes. If you can update this thread when the fix is rolled into a future firmware update, that would be appreciated!

[Anonymous user]

Confirmed this is working in v6.5.0

[Terry_Kennedy_apc]

On 7/30/2018 3:52 PM, Angela said:

We have a 6.6.4 coming in a few weeks but I think it will be too late in the development process to fix it in that rev unfortunately but I'll see - just wanted to set expectations.

It is still broken as of 6.7.2 (but not noted in the "known problems" section of the release notes). Can you bug someone to make sure it gets fixed in the next release? Also, it would be nice if a test for this got added to the QA testing - this regression has come back a couple of times after being fixed.

[BillP]

Hi Terry,

I have added your comments to the engineering notes. The issue is listed to be address in a future release.

[Terry_Kennedy_apc · Accepted solution]

Thanks!