nut-debian/docs/man/upsdrvsvcctl.txt
2022-07-10 09:23:45 +02:00

206 lines
7.3 KiB
Plaintext

UPSDRVSVCCTL(8)
===============
NAME
----
upsdrvsvcctl - UPS driver service instance controller
SYNOPSIS
--------
*upsdrvsvcctl* -h
*upsdrvsvcctl* ['OPTIONS'] {start | stop } ['ups']
DESCRIPTION
-----------
*upsdrvsvcctl* provides a uniform interface for controlling your UPS
drivers wrapped into service instances on platforms which support that
(currently this covers Linux distributions with systemd and systems
derived from Solaris 10 codebase, including proprietary Sun/Oracle
Solaris and numerous open-source illumos distributions with SMF).
It may be not installed in packaging for other operating systems.
When used properly, upsdrvsvcctl lets you maintain identical startup
scripts across multiple systems with different UPS configurations.
The goal of this solution is to allow the services of *upsd* data
server to start up even if some of the power devices are currently
not accessible, and for NUT drivers to be automatically restarted
by the system in case of problems (driver bug, startup failure).
It also allows for faster startup of systems which monitor several
devices, by letting each driver to start in parallel with others,
and not with a sequential loop like was done previously.
Independent service instances for each NUT driver also allow one
to configure further dependencies, such as that networking must be
available for SNMP and similar drivers (but is not needed for
local-medium drivers such as serial or USB).
The old monolithic "all or nothing" solution requiring that all
drivers must be running, which sufficed for deployments with a few
UPSes, did not really work well for monitoring larger deployments.
It was also not easy to strike a pre-packaged balance between early
UPS protection for USB/serial home setups vs. waiting for network
on larger ones.
*upsdrvsvcctl* is a script which mimics the operation of *upsdrvctl*
program (where possible) to provide similar end-user experience when
manipulating drivers wrapped into service instances rather than as
directly executed daemons. It relies on *nut-driver-enumerator.sh*
for a large part of actual operations.
You should use upsdrvsvcctl instead of direct calls to the drivers
and daemon-based management with *upsdrvctl* whenever possible (that
is, for "production" use on compatible OSes). Otherwise (testing,
other OSes) the *upsdrvctl* is a recommended option.
OPTIONS
-------
*-h*::
Display the help text.
*-t*::
Enable testing mode. Testing mode makes upsdrvsvcctl display the actions
it would execute without actually doing them.
OPTIONS OF UPSDRVCTL NOT (CURRENTLY) APPLICABLE TO UPSDRVSVCCTL
---------------------------------------------------------------
Options like '-r', '-u' or '-D' could be handled by properties of the
service instances themselves, with this script helping to configure
them (assuming proper privileges of the user who called it). This is
not a "production" use case, though, to change such options on a
configured system -- so for experiments and troubleshooting, it may
be better to stop the service instance and play with *upsdrvctl*
directly.
*-r* 'directory'::
If starting a driver, this value will direct it to *chroot*(2) into
'directory'. This can be useful when securing systems.
This may be set in the ups.conf with "chroot" in the global section.
*-u* 'username'::
If starting a driver, this value will direct it to *setuid*(2) to
the user id associated with 'username'.
If the driver is started as root without specifying this value, it will
use the username that was compiled into the binary. This defaults to
"nobody", and is far from ideal.
This may be set in ups.conf with "user" in the global section.
*-D*::
Raise the driver debug level. Use this multiple times for additional
details.
COMMANDS
--------
*upsdrvsvcctl* supports three of the commands processed by *upsdrvctl* --
start, stop and shutdown. They take an optional argument which is a UPS
name from linkman:ups.conf[5]. Without that argument, they operate on
every UPS that is currently configured.
Note: shutdown is currently supported by stopping the driver service
instances to release the potentially held ports etc., calling the
*upsdrvctl* directly for issuing the shutdown command, and restarting
the driver service instances to reconnect when the device comes back
online.
*start*::
Start the UPS driver(s). In case of failure, further attempts may be executed
by using the 'maxretry' and 'retrydelay' options - see linkman:ups.conf[5].
*stop*::
Stop the UPS driver(s).
*upsdrvsvcctl* also supports further operations for troubleshooting the
mapping of NUT driver section names to the service instance names (which
may differ due to limitations of various systems).
*list*::
list the currently active mapping of service instances to device sections
*resync*::
update the mapping of service instances for NUT drivers to device section
names used in 'ups.conf' (register new instances, tear down obsoleted ones).
COMMANDS OF UPSDRVCTL NOT (CURRENTLY) APPLICABLE TO UPSDRVSVCCTL
----------------------------------------------------------------
*shutdown*::
Command the UPS driver(s) to run their shutdown sequence. Drivers are
stopped according to their sdorder value - see linkman:ups.conf[5].
WARNING: this will probably power off your computers, so don't
play around with this option. Only use it when your systems are prepared
to lose power.
NOTE: refer to linkman:ups.conf[5] for using the *nowait* parameter.
ENVIRONMENT VARIABLES
---------------------
*NUT_CONFPATH* is the path name of the directory that contains
`upsd.conf` and other configuration files. If this variable is not set,
*upsdrvsvcctl* (or rather *nut-driver-enumerator.sh*) would use a built-in
default, which is often `/usr/local/ups/etc`.
DIAGNOSTICS
-----------
upsdrvsvcctl will return a nonzero exit code if it encounters an error
while performing the desired operation. This will also happen if a
driver takes longer than the 'maxstartdelay' period to enter the
background.
Any messages issued by the *upsdrvctl* program used to start the NUT
drivers as part of the service instances' implementations, or by the
drivers themselves, will be logged by the service management framework
facilities and will not appear in your interactive terminal used to
manage the driver.
Use `upsdrvsvcctl list` or `upsdrvsvcctl list NUT-device` to find out
the service instance name for the NUT driver (section name) you are
interested in. Then look up the service logs (where the outputs of the
service implementation program as well as the framework messages about
this service are stored), as suggested below:
*Linux systemd*::
Messages will normally be kept in the service journal, so:
journalctl -lu nut-driver@instance-name
Note that your local system configuration may be impacted by such
nuances as passing the journal data to a standard syslog server,
and/or by having a small cache for locally stored journal messages
(so older entries would disappear). There may also be or not be a
copy of the journals stored in the filesystem.
*Solaris SMF*::
Look for `/var/svc/log/system-power-nut-driver:instance-name.log` file.
AUTHOR
------
Jim Klimov <jimklimov+nut@gmail.com>
SEE ALSO
--------
linkman:upsdrvctl[8], linkman:nutupsdrv[8], linkman:upsd[8],
linkman:nut-driver-enumerator[8], linkman:ups.conf[5]
Internet resources:
~~~~~~~~~~~~~~~~~~~
The NUT (Network UPS Tools) home page: http://www.networkupstools.org/