BG500 network failure

[svhann_apc]

I purchased 2 new BG500 UPSs.  After working a couple weeks, both now fail to acquire addresses by DHCP.  As you would expect, with no address they thrash the devices connected to the smart outlets using the watchdog feature.

Do I have 100% failure or could there be something I missed?

FYI, the current firmware for these devices is not posted on APC's web site, but they will send you a link if you ask.

[BillP]

Hmm, this does not sound right. The network management itself has a watchdog feature to only reboot it (and NOT the outlets) if it can't see any network traffic and has a suspicion its TCP/IP stack needs to be rebooted.

Can you directly connect into the device via a serial cable or something so we can look at it (since it can't get DHCP)? I am interested in an event log.

I think this device has only had one firmware release if I am not mistaken so the original may not be posted..

I feel like we are missing something for sure that hopefully we can figure out. I was thinking, the regular network management cards (which this device has a mini version inside) have a pinhole reset button you can hold down to "format" the device and erase all of the settings inside it (including any corruption). You hold it down for 20+ seconds and it is supposed to format/wipe the device. Could try that too.

[svhann_apc]

When the first BG500 (192.168.1.189) failed I purchased a second BG500 (192.168.1.188) and had someone install it.  It had different firmware. APC Tech Support sent me a link to the updated firmware.  I updated the first one to the new firmware today.

I'm travelling and need  my home network to stay up.  That's why I bought this UPS.  Pushing the network interface reset button is not a useful option.

When I got home, I removed the second BG500 and reinstalled my old  BH500NET.  I needed to stop the BG500 from constantly resetting my network gear.  Today I had time to reconnect both BG500s to the network.  They both acquire addresses.

Looking at the logs (attached) I saw both BG500s were getting IPv6 addresses.  This is the default configuration. I turned off IPv6 on both units today.

My plan is to put both of them on the network without anything attached to the outlets.  I will clear both event logs.  I can watch for a couple days and see if anything happens.  If not, I can simulate a ping failure by pointing them to a unused local address and see what happens.  Perhaps I can capture the failure in an event log.

[BillP]

Config.ini file may help us too. In both logs, I see this:

06/23/2015 06:21:13 Device UPS: The power for Smart Outlet 2 [Smart Outlet 2] is on. 0x0134
06/23/2015 06:20:14 Device UPS: The power for Smart Outlet 1 [Smart Outlet 1] is on. 0x0134
06/23/2015 06:20:05 Device UPS: The power for Smart Outlet 2 [Smart Outlet 2] is off. 0x0135
06/23/2015 06:20:05 Device UPS: The power for Smart Outlet 1 [Smart Outlet 1] is off. 0x0135
06/23/2015 06:20:03 Device UPS: Smart Outlet 2 [Smart Outlet 2] has been commanded to reboot, sequenced. 0x0174
06/23/2015 06:20:03 Device UPS: Smart Outlet 1 [Smart Outlet 1] has been commanded to reboot, sequenced. 0x0174

So according to the log, these cycles on and off are logging themselves as reboot commands. I was wondering if something (that is not you) is somehow commanding this to reboot. On larger UPSs that use the same management interface, we sometimes see stuff doing this on the SNMP interface but I believe this UPS does not support SNMP. Config.ini may give me a clue if there is something set improperly or an interface enabled that could be somehow erroneously accepting commands or something.

Also, IPv6 and IPv4 are on by default. Disabling IPv6 is OK and makes sense if you're not using it but the only IPv6 address it was getting was an auto-assigned one that should have no affect on the unit's operation or communication.

And as you mentioned, I do see now that the AOS 6.0.1 and Back-UPS app v6.0.4 are the latest.

[svhann_apc · Accepted solution]

Config.ini files attached.

When I configured the UPS, I set a permanent DHCP reservation in the router for the IP addresses.

During the time each  UPS was resetting, I remotely logged into my router.  The UPS was not listed as an attached device.  I was unable to ping the UPS or connect to the network interface of the UPS.  Since there was no network connection, I couldn't do anything else with the UPS. I did not try to use the Device IP Configuration Utility for off network devices.

I reached the conclusion the issue was -- since the UPS had no valid address, it could not successfully ping the cable modem. Since the ping failed, the UPS assumed the cable modem was down and cycled the smart outlets.

APC tech support suggested I assign static IP addresses.  I did that with the replacement UPS and put it back online.  The original UPS is configured as it was and is connected to the network.  According to the event logs, neither UPS has cycled since July 24th. According to the router, both are connected to the network.