| commit | author | age | ||
| 2313e2 | 1 | This is the directory for Network UPS Tools configuration files, |
| b4f492 | 2 | as distributed by OpenIndiana Hipster. |
| 2313e2 | 3 | |
| b4f492 | 4 | Some example files are provided as `/usr/share/nut/etc-samples/*.sample`. |
| AP | 5 | Actual configuration files are loaded by programs from `/etc/nut`. |
| 6 | General workflow is to modify provided sample configuration files | |
| 7 | `/etc/nut/*.conf` as fits your situation. | |
| 2313e2 | 8 | |
| JK | 9 | NOTE: Take care about proper ownership and access bits of configuration |
| 10 | files, especially for those with credentials. | |
| 11 | ||
| 12 | On a NUT server machine that directly tracks UPS states, you must | |
| 13 | configure: | |
| 14 | * `ups.conf` with a block for each driver configuration (required) | |
| 15 | * `upsd.users` with a block for each set of credentials and allowed | |
| 16 | actions so different clients can do different things (required for | |
| 17 | monitoring, `upsrw` CGI and other clients) | |
| 18 | * `upsd.conf` can customize `upsd` server settings (optional) | |
| 19 | ||
| 20 | On a NUT client with `upsmon` for graceful shutdown (may be or not be | |
| 21 | the same as teh NUT server machine), you must configure: | |
| 22 | * `upsmon.conf` with a `MONITOR` line for each monitored UPS, including | |
| 23 | how many power-supplies of this server it feeds, and a `MINSUPPLIES` | |
| 24 | line to define the critical threshold (how many power-supplies should | |
| 25 | have stable external power conditions for the system to not start its | |
| 26 | graceful shutdown when UPS battery goes critical). Note that an admin | |
| 27 | workstation can monitor lots of devices with none (0) feeding it, just | |
| 28 | to aggregate notifications about state changes, etc. in one place. | |
| 29 | There are also other settings to configure for communication frequency, | |
| 30 | shutdown and notifications. | |
| 31 | ||
| 32 | On a NUT client with `upssched` you can configure complex reactions to | |
| 33 | change in UPS state, including a way to ignore occasional short-lived | |
| 34 | state flip-flops. The daemon tracks the beginning of some state and sets | |
| 35 | a user-defined delayed timer to react to the state later - unless the | |
| 36 | state is cleared before the timeout. To use this facility, you must | |
| 37 | configure: | |
| 38 | * `upssched.conf` with settings for events and reactions | |
| 39 | * `upssched-cmd` is a sample script with a `case` structure to handle events | |
| 40 | ||
| 41 | For CGI clients you should set up the web-server of your choice accordingly | |
| b4f492 | 42 | to run the binaries provided in /usr/lib/nut/cgi-bin, and configure: |
| 2313e2 | 43 | * `upsset.conf` contains a flag that you must set to assert that you |
| JK | 44 | have indeed configured and enabled CGI clients knowingly, and have set |
| 45 | up proper web-server security (required) | |
| 46 | * `hosts.conf` contains a list of `MONITOR`ed devices that the CGI clients | |
| 47 | would display and/or send commands to (required) | |
| 48 | * `upsstats.html` and `upsstats-single.html` are templates for rendering | |
| 49 | the CGI web-interface regarding all known, or a specific chosen, devices. | |
| 50 | ||
| 51 | Beside directly NUT configuration examples, this distribution may include: | |
| b4f492 | 52 | `/usr/share/nut/etc-samples/avahi-nut.service.sample` with settings for |
| AP | 53 | mDNS announcement of NUT service on the Avahi zero-conf subnet. |
| 54 | This sample may be imported into your Avahi installation. | |
| 2313e2 | 55 | |
| JK | 56 | Hope this helps, |
| 57 | //Jim Klimov | |
