draidd(1M) draidd(1M)
NAME
draidd - a daemon that monitors both channels of a RAID box
SYNOPSIS
draidd [[-D [-cCru] [-M number]]
draidd [-gGHsS]] [-lLv] [-x number] [-t file] [-f configuration-file]
[config-device]
draidd [-a] [{-e|-d|-h}] draid-device
DESCRIPTION
draidd (Dual RAID Channel Daemon) is usually started by means of
init(1M) when the system is booted. draidd reads its configuration
information from the configuration-file at startup; the default name
for this file is /etc/draid. See also draid(4) for the format of the
file.
draidd creates a redundant access to a RAID box with two controllers
which are connected to two host controllers. The daemon uses a system
driver which allows access to the RAID box over one or both controll-
ers.
If a controller cannot be accessed, draidd reconfigures the routing
table of the system driver, and consequently all further access to the
RAID box takes place over the other controller.
draidd rereads its configuration file as soon as it receives a hangup
signal (SIGHUP). New RAID box configurations can thereby be activated
and existing RAID box configurations can be deleted or modified. The
new configuration can be used as soon as the RAID box has been config-
ured successfully.
OPTIONS
draidd can be started as a daemon (-D option) or draidd can be used to
read the actual configuration.
The following options can only be used if draidd should be started as
a daemon:
-C Uses and configures the RAID box, as described in the
configuration-file.
-c Uses the current configuration of the connected RAID box
(default) until a SIGHUP signal is received.
-D Becomes a daemon. This flag must only be specified for the draidd
process which is intended to monitor the RAID boxes. Only one
process per system may be started using this flag.
Page 1 Reliant UNIX 5.44 Printed 11/98
draidd(1M) draidd(1M)
-M number
Sets the maximum number of reconfiguration attempts (default: 1).
Once the maximum number of reconfiguration attempts has been
reached, access to the entire RAID box is deactivated.
-r Creates a "SCSI bus reset" before reconfiguring a RAID box,
should a controller fail.
It is recommended that you do not initiate a "bus reset". This
flag should only be used in cases where it is clear what is meant
by a "bus reset". Data for other SCSI devices on this bus may be
lost.
-u Updates the autoconf-tree after reconfiguration. By default this
is switched off. If -u is set a reconfiguration may take a longer
period of time.
The following options can only be used if draidd should not be started
as a daemon:
-G Displays the RAID box configuration in detail.
-g Displays the configuration of the RAID box.
-H A hangup signal (SIGHUP) is sent to the active daemon.
-S Displays statistical data, also for asynchronous tasks.
-s Displays statistical data.
Options for high-availability configurations:
A special draid device (e. g. draid0) must always be nominated for
these options.
-a Log information is written to a file on completion of all activi-
ties. The file name is made up of the name of the draid device
and the process ID.
-h The specified draid device is reconfigured.
The following options are only supported by EMC RAID systems.
-e The SCSI commands sent to the ports of the RAID systems write
protect the LUNs of the port to prevent them being accessed from
ports other than those connected to this host.
-d Access to the LUNs of the draid device is locked.
Page 2 Reliant UNIX 5.44 Printed 11/98
draidd(1M) draidd(1M)
The following options are ever valid:
-f configuration-file
Uses the configuration-file as a configuration description.
(default: /etc/draid) [see draid(4)].
-L Output from draidd is not logged with syslogd(1M), but is sent to
the system console. This is the default setting.
-l Logs output from draidd using syslogd(1M). This option may cause
a shutdown in high-load systems when an error occurs, because a
file system can no longer be accessed.
-t file
Instead of using syslogd(1M) the file file will be used for the
programs output. This is useful, if the output of draidd starts
very early during system startup, where syslogd(1M) is not avail-
able.
-v Outputs the program segment where draidd is found. The volume of
output increases with each occurrence of -v. syslogd(1M) is used
to output the information.
-x 1|2|3
Sets the debug level for the system driver. This should only be
used for testing or debugging.
config-device
To enable the required configuration to be loaded in the system
driver [see draid(4)], the device nodes can be specified as
config-device. If no config-device is specified during configura-
tion, /dev/draid/draid0l0s0 can be used. This device node is only
used to enable communication with the system driver; the nodes
are configured as described in draid(4).
This device node must be specified as config-device to enable
reading-in of the statistical data of a device node.
SIGNALS
draidd can be affected during runtime by the issuing of signals. Apart
from the HUP signal, all other signals should only be used for debug-
ging or testing.
SIGHUP draidd reconfigures the RAID box. -H option.
SIGUSR1 Output is issued indicating the program part in which draidd
is contained. The volume of output is increased every time
SIGUSR1 is sent. -v option.
SIGUSR2 The volume of output is reduced every time SIGUSR2 is sent.
-v option.
Page 3 Reliant UNIX 5.44 Printed 11/98
draidd(1M) draidd(1M)
SIGPWR Enables or disables the use of syslog(3C) for output. -l or
-L options.
SIGPROF draidd logs information in an internal buffer. When the
SIGPROF signal is issued, this buffer is output on the sys-
tem console and written to the /tmp/draidd.dump file.
FILES
/etc/draid
Configuration description.
/dev/draid/draid0
This device node is used by draidd to communicate with the
driver.
/tmp/draidd.dump
Internal log information of draidd daemon.
./draidX.PID
Internal log information of a draidd process; X is the number of
affected draid devices.
/dev/autoconf
This device node is used to determine the RAID controllers device
addresses.
/dev/draid
Device node directory.
/tmp/.draiddlock
draidd never creates this file, however it does delete it follow-
ing reconfiguration. This file can be created and tested using an
external configuration process.
SEE ALSO
draid(4), draid(7).
Page 4 Reliant UNIX 5.44 Printed 11/98