Configuring UPS/apcupsd

With my new server I bought a UPS, partly just as a Good Thing, but also because I suspect a powercut fried the motherboard on a previous machine that I had, and this baby is too precious to lose ;)

The idea is that the UPS will smooth out the power supply to my server, protecting it from surges or temporarily blips in power loss. If there's a proper power cut, the UPS is connected to my server and can initiate a graceful shutdown instead of system crash. It seems unintuitive in this day and age of laptops and iPads that you just close or switch off to "suspend" them that killing the power to a server can damage it, but when you think about it just a moment more, it's hardly surprising.

The software for enabling the server (running Debian 8/Proxmox 4, kernel 4.4.6-1-pve), is a textbook example of why many people took against Linux in the early days - esoteric, out of date documentation. In this short article I'll record how I actually got the thing to work, and a couple of errors I hit along the way.

The UPS I'm running is a APC Back-UPS 650, which includes a cable to connect its data port to the USB of a server/workstation

Detour - nuts

My first port of call was the top of the Google hits, nut (Network UPS Tools), using the usbhid-ups driver. This failed to even start up:

usbhid-ups[11055]: segfault at 0 ip 00007f4e9fe3ca87 sp 00007fffcbd48228 error 4 in[7f4e9fd1d000+1a2000]

Success - apcupsd

Next I found apcupsd (APC UPS Daemon).

apt-get install apcupsd  

Finding the device

The docs suggested I'd find my UPS device listed at /proc/bus/usb/drivers, and if not the drivers at least at /proc/bus/usb/drivers - neither of these worked for me:

root@proxmox01:/proc/bus# cat /proc/bus/usb/devices  
cat: /proc/bus/usb/devices: No such file or directory  
root@proxmox01:/proc/bus# cat /proc/bus/usb/drivers  
cat: /proc/bus/usb/drivers: No such file or directory  

But a fortunate Google hit reminded me about lsusb:

root@proxmox01:/proc/bus# lsusb  
Bus 003 Device 007: ID 051d:0002 American Power Conversion Uninterruptible Power Supply  

And there it is. Phew. It's also found using udev (per the apcupsd documentation - maybe I shouldn't be quite so rude about it):

root@proxmox01:/proc/bus# udevadm info --attribute-walk --name=/dev/usb/hiddev0  
  looking at parent device '/devices/pci0000:00/0000:00:14.0/usb3/3-2':
    ATTRS{bNumInterfaces}==" 1"
    ATTRS{serial}=="4B1517P07342  "
    ATTRS{version}==" 1.10"
    ATTRS{manufacturer}=="American Power Conversion"
    ATTRS{product}=="Back-UPS CS 650 FW:817.v9.I USB FW:v9"

And finally, it's also there under usb-devices:

root@proxmox01:/proc/bus# usb-devices  
T:  Bus=03 Lev=01 Prnt=01 Port=01 Cnt=02 Dev#=  7 Spd=1.5 MxCh= 0  
D:  Ver= 1.10 Cls=00(>ifc ) Sub=00 Prot=00 MxPS= 8 #Cfgs=  1  
P:  Vendor=051d ProdID=0002 Rev=00.06  
S:  Manufacturer=American Power Conversion  
S:  Product=Back-UPS CS 650 FW:817.v9.I USB FW:v9  
S:  SerialNumber=4B1517P07342  
C:  #Ifs= 1 Cfg#= 1 Atr=e0 MxPwr=0mA  
I:  If#= 0 Alt= 0 #EPs= 1 Cls=03(HID  ) Sub=00 Prot=00 Driver=usbhid  

Found the device, now configure apcupsd

In /etc/apcupsd/apcupsd.conf I set:


I started the daemon, which didn't actually run based on the status command:

root@proxmox01:/proc/bus# service apcupsd start  
root@proxmox01:/proc/bus# service apcupsd status  
● apcupsd.service - LSB: Starts apcupsd daemon
   Loaded: loaded (/etc/init.d/apcupsd)
   Active: active (exited) since Fri 2016-06-17 13:56:25 BST; 28min ago

Jun 17 13:56:25 proxmox01 apcupsd[15897]: Please check your configuration ISCONFIGURED in /etc/default/apcupsd  

After editing /etc/default/apcupsd to set


The service came up and stayed up:

root@proxmox01:/proc/bus# service apcupsd restart  
root@proxmox01:/proc/bus# service apcupsd status  
● apcupsd.service - LSB: Starts apcupsd daemon
   Loaded: loaded (/etc/init.d/apcupsd)
   Active: active (running) since Fri 2016-06-17 14:30:01 BST; 1s ago
  Process: 1242 ExecStop=/etc/init.d/apcupsd stop (code=exited, status=0/SUCCESS)
  Process: 1249 ExecStart=/etc/init.d/apcupsd start (code=exited, status=0/SUCCESS)
   CGroup: /system.slice/apcupsd.service
           └─1254 /sbin/apcupsd

Jun 17 14:30:01 proxmox01 apcupsd[1249]: Starting UPS power management: apcupsd.  
Jun 17 14:30:01 proxmox01 apcupsd[1254]: apcupsd 3.14.12 (29 March 2014) debian startup succeeded  
Jun 17 14:30:01 proxmox01 apcupsd[1254]: NIS server startup succeeded  

The successful service start can also be seen in /var/log/ and /var/log/daemon.log

Testing it

Now we're up and running, let's test it.

Disconnect UPS data cable from Server

The unscary one first - disconnect the data comms between UPS and server. It took a few more than the 6 seconds than the doc says, but within a minute or so I got a system beep on the server and :

Broadcast message from root@proxmox01 (somewhere) (Fri Jun 17 14:39:48 2016):

Warning communications lost with UPS proxmox01  

as well as an entry in the /var/log/daemon.log:

2016-06-17 14:39:48 +0100  Communications with UPS lost.  

I plugged the USB back in to my server, and got:

Jun 17 14:41:14 proxmox01 apcupsd[1254]: Communications with UPS restored.  

Disconnect UPS from Mains Power

The document entertainingly says:

The first time that you do this, psychologically it won't be easy, but after you have pulled the plug a few times, you may even come to enjoy it.

The first step is to replace the actual script that would shut the server down if necessary on power failure and battery exhaustion with a dummy one:

mv /etc/apcupsd/apccontrol /etc/apcupsd/apccontrol.bak  
cp /usr/share/doc/apcupsd/examples/safe.apccontrol /etc/apcupsd/apccontrol  

And then ... pull the plug on the USB from the mains ...

Broadcast message from root@proxmox01 (somewhere) (Fri Jun 17 14:49:41 2016):

apccontrol: Warning power loss detected.

Broadcast message from root@proxmox01 (somewhere) (Fri Jun 17 14:49:47 2016):

apccontrol: Power failure. Running on UPS batteries.  

Entries in the /var/log/daemon.log, and my UPS starts beeping too.

Plugging the UPS back in (phew), and:

Broadcast message from root@proxmox01 (somewhere) (Fri Jun 17 14:51:04 2016):

apccontrol: Off battery. Mains returned.

Broadcast message from root@proxmox01 (somewhere) (Fri Jun 17 14:51:04 2016):

apccontrol: Power has returned...  

In real life when we had a powercut the UPS worked perfectly too:

Jul 18 02:19:00 proxmox01 apcupsd[1254]: Power failure.  
Jul 18 02:19:06 proxmox01 apcupsd[1254]: Running on UPS batteries.  
Jul 18 02:25:40 proxmox01 apcupsd[1254]: Mains returned. No longer on UPS batteries.  
Jul 18 02:25:40 proxmox01 apcupsd[1254]: Power is back. UPS running on mains.  

Full Power Outage


And yes this is bad. Just like backups being only as good as the last successful restore, a UPS graceful shutdown routine really does need to be tested. Watch this space!

Bonus - data extract

Because we left NETSERVER enabled in the config, we can probe the stats of the UPS:

root@proxmox01:/proc/bus# apcaccess status  
APC      : 001,045,1060  
DATE     : 2016-06-17 14:30:35 +0100  
HOSTNAME : proxmox01  
VERSION  : 3.14.12 (29 March 2014) debian  
UPSNAME  : proxmox01  
CABLE    : USB Cable  
DRIVER   : USB UPS Driver  
UPSMODE  : Stand Alone  
STARTTIME: 2016-06-17 14:30:01 +0100  
MODEL    : Back-UPS CS 650  
LINEV    : 244.0 Volts  
LOADPCT  : 40.0 Percent  
BCHARGE  : 100.0 Percent  
TIMELEFT : 14.9 Minutes  

I might brew a bash-based extract of this data into InfluxDB to track in Grafana (of course!), or maybe look at this custom Telegraf plugin.

Robin Moffatt

Read more posts by this author.

Yorkshire, UK