"Communication Lost" Smart UPS 1000 W2K3 R2

[BillP]

Hi all,

We have a large number of W2K3 R2 servers at remote sites hooked up to Smart UPS 1000s by USB. They seem to have a tendency to lose communication for no apparent reason and not regain it.
The driver is still present, and reinstalling it succeeds with no effect. Restarting the agent service has no effect. I neglect to mention the server or console as the web interface provided by the agent also reports comms lost.

I should mention we are having exactly the same problem with an SUA3000RM on another W2K3 server too.

Any thoughts? Obviously this is a big issue as we cannot easily gain physical access to the equipment.

Running 7.0.5.

Cheers
Rob

[BillP]

Two different models of UPS, but about 5 or 6 different physical units exhibiting the same issue.

I apologise for my disobedience, but these UPS' are all running critical systems, and are also several miles away.

[TheNotoriousKMP_apc]

Rob,

Have you tried to brain-dead the UPS and fully enable the USB support on the UPS?

To do so:

Power off the load attached to the UPS
Power off the UPS and disconnect from the wall
Remove any communications cables
Depress the Off button for 4-6 seconds until you hear a click.

Load everything back up and keep an eye on it.

[BillP]

I haven't, however as I mentioned this applies to multiple different units, and I don't therefore see that it could be a one-off fault fixed by what is effectively rebooting the unit.

Worthy of note that if I wanted to re-establish USB comms I could disconnect the USB cable and reconnect it. However this will not fix the re-occurring issue.

[TheNotoriousKMP_apc]

You have two different UPS's both of which have serial and USB support. I would try that first.

[TheNotoriousKMP_apc · Accepted solution]

I apologize if that previous post came off as touchy.

I understand that it's a critical setup, as is the case we run into all the time doing support. There usually is a requirement of some type of downtime window that most customers receive to do troubleshooting.

The only reason I threw it out there is because it seems to be a universal problem, and is more than just a coincidence that it's occurring on two different systems that both have the ability to have either one or the other means of communication.