Tripp Lite UPS keeps disconnecting/reconnecting

[superwad]

I'm just about to move my server into its final resting place after doing a 2-month long live test of everything. I'm pretty sure I have all the kinks worked out, save this one.

I have a Tripp Lite SMART1500LCD 1500VA 900W UPS hooked up to the system via USB, and it seems that every day I get an email that contains some amount of this:

Code:

> ugen1.3: <Tripp Lite> at usbus1 (disconnected)
> uhid0: at uhub3, port 1, addr 3 (disconnected)
> ugen1.3: <Tripp Lite> at usbus1
> uhid0: <Tripp Lite TRIPP LITE UPS, class 0/0, rev 1.10/0.10, addr 3> on usbus1
> ugen1.3: <Tripp Lite> at usbus1 (disconnected)
> uhid0: at uhub3, port 1, addr 3 (disconnected)
> ugen1.3: <Tripp Lite> at usbus1
> uhid0: <Tripp Lite TRIPP LITE UPS, class 0/0, rev 1.10/0.10, addr 3> on usbus1
> ugen1.3: <Tripp Lite> at usbus1 (disconnected)
> uhid0: at uhub3, port 1, addr 3 (disconnected)
> ugen1.3: <Tripp Lite> at usbus1
> uhid0: <Tripp Lite TRIPP LITE UPS, class 0/0, rev 1.10/0.10, addr 3> on usbus1
> ugen1.3: <Tripp Lite> at usbus1 (disconnected)
> uhid0: at uhub3, port 1, addr 3 (disconnected)
> ugen1.3: <Tripp Lite> at usbus1
> uhid0: <Tripp Lite TRIPP LITE UPS, class 0/0, rev 1.10/0.10, addr 3> on usbus1
> ugen1.3: <Tripp Lite> at usbus1 (disconnected)
> uhid0: at uhub3, port 1, addr 3 (disconnected)

I haven't yet set up my UPS on the system; I was waiting until I got closer to D-Day to finish that task, but it looks like that's going to be this weekend or next at the latest, so now would be the time to sort this out.

As I haven't done any troubleshooting on this disconnect issue specifically other than look at these daily emails and archive them, what would my first steps be? The UPS actually works properly (I've run the self-diagnostic on it), and I have noticed zero power issues with the computer since adding this UPS. There are multiple USB ports, and I've tried different ones (USB 2.0 and 3.0) and the problem persists (granted, I haven't tried every port on the board).

I did try to set up the UPS service once, but the service didn't start; I'm going to assume that was because I hadn't actually set up the config file but was just trying to start it on whatever the default config settings were.

Any help?

 

[superwad]

Haven't received a reply in ~1 week. Would still like some insight into if this is a problem or not, and how I can eliminate the disconnects.

 

[errmatt]

Had this exact problem with this exact model of UPS. Fought with it for several months trying to make it report and reliably shut down the FreeNAS server. I finally gave up and swapped it for an older APC unit with less capacity that I had on the shelf. FreeNAS now detects that UPS and shuts down properly and reports properly. I even get e-mails when it goes on battery and when it comes back off battery. Less capacity, but at least I can rely on it to shut down properly if I'm not there when the power goes out.

I've since tried to use the same Tripp Lite UPS on a physical windows box..... I've given up on making it work/report properly to the OS on that machine as well, since half the time I boot it or reboot it, it shuts down immediately after booting into Windows because the UPS is reporting being at 0% capacity. At this point, my Tripp Lite UPS is no longer under warranty and I've given up on relying on it to report properly to any machine, so I now only use it to plug my switch and router and security camera system into, since those don't require proper shutdowns, they just run until the battery gives out.

So, I don't know if it's faulty or if it's just a crappy product. If yours is under warranty, I would try to get Tripp Lite's support on the hook for fixing or replacing it. As for me, I'll stick with APC for UPS's from now on. Sorry I don't have a better answer for you. Try it on a Windows box, I bet you'll run into the same problem I had.

 

[superwad]

If that's the case, it's sadly too late for me to return it (by about 6 weeks from Amazon). I could probably try the warranty route, but it might be just that the product is entirely defective and they can't do anything about it.

Do you have recommendations for a similarly powered UPS that's still rack mountable?

 

[Fox]

I have this exact problem. I have contacted the USB mailing list for FreeBSD, and they *may* have a fix.. I also have another thread:
https://forums.freenas.org/index.php?threads/usb-kernel-disconnect-messages-filling-up-log-ups.25677

I will know more shortly.. I am still in the middle of the process..

 

[Fox]

Got the fix (supposedly), from FreeBSD programmer. He took pity on me and it looks like he coded it. I hope it works.

I entered a feature request in FreeNAS (listed below).. Please add your name to it to show support, or respond to the threads so the developers know this is a desired fix and hopefully raise it on the priority list. I would love to have my UPS actually working for Christmas. I hate having to shut down my FreeNAS box every time I leave for an extended period of time.

Bug/Feature Request:
https://bugs.freenas.org/issues/7153

 

[superwad]

Awesome. I'll add my name to the bug! Thanks for taking the initiative on getting this resolved Fox :)

 

[Fox]

As I mentioned in the bug report and the other thread, I was unsuccessful in getting it to work. The fix as it turned out was for another unrelated USB issue.. :-(

So, what I did do as an initial work around was to purchase a cheaper UPS (CyberPower) and hook it up to my NAS. It was one of the Amazon lightning deals, which offers a 600W UPS with pure Sine Wave for $99. I wanted one for my workstations, so I thought I would grab it. It works for the FreeNAS box as well, but doesn't give me as much time to shutdown, and it is not actively cooled like the Tripp Lite. If I want the higher power of the Tripp Lite, I need to plug my FreeNAS power cable into it, and then plug the Cyberpower into the USB port of the NAS, so that the Cyber Power will trigger a FreeNAS shutdown when the power goes out (but it won't provide power to the FreeNAS). This is just a hack, but it does solve the problem in the short run (but not the long run).

I think my next step is to load the full FreeBSD version, and start seeing if I can trouble shoot the issue. I have now a working system, and a non-working one, and the ability to compare the results (usbdump), so perhaps I can figure out what the issue is. I also provided more input to the mailing list (usbdump output), but I am guessing it might take a while before they can look at it.

I also want to test the latest version of FreeBSD to see if any changes may have already fixed the issue.

Unfortunately, I am very busy, so this may take a while.