STATES AND ACTIONS OF RUN CONTROL PROXIES
The list of States and Actions for each proxy define the state model
that the proxy presents to the central run control and specify the
interface between the two.
Database Proxy
Process name = database
Object name = <partition>::DATABASE
Provides interface for all database interaction, both information retrieval and logging, whether or not a real database is used. A separate proxy is required for each partition. By default the Objectivity configuration and conditions
databases are accessed, a flat file fall back is provided by the database_noobbjy proxy.
States and Actions:
state action
ERROR
RESET
READY
SET_RUN_NUMBER (run_number)
GET_RUN_NUMBER
GET_CONFIGURATION_KEY (run_type)
GET_CRATE_MASK (run_type)
LOG_CONFIGURE (runtype, pname, pnumber, key, crate_mask, trigger_mask)
LOG_UNCONFIGURE (runtype, pname, pnumber, key, crate_mask, trigger_mask)
LOG_BEGIN (run_number, partition_number, runtype, pname, key, trigger_mask, crate_mask, logging)
LOG_END (run_number, partition_number, runtype, pname, key, trigger_mask, crate_mask, logging)
REFRESH
SHUTDOWN
PROXY_DEAD
Action details:
SET_RUN_NUMBER (run_number)
This action is no longer supported and doea nothing. It will be removed.
GET_RUN_NUMBER
Sets the object parameter DB_RUN_NUMBER to the next valid run number. The last run number is read either from a file (default, file name = RUNNUMBER) or from Objectivity and incremented by 1. The new number is then stored.
GET_CONFIGURATION_KEY (run_type)
Sets the object parameter DB_KEY to the current valid key for the given run type. The key <-> runtype mapping is either read from a file (default, file name = Orc/runtype.dat) or from Objectivity (takes most recent valid TopMap from the
configuration database).
LOG_CONFIGURE (runtype, partition_name, partition_number, key, crate_mask, trigger_mask)
Logs the given information to cout and to the database as a configure record (OrcConfigRecordP in Objectivity).
LOG_UNCONFIGURE (runtype, partition_name, partition_number, key, crate_mask, trigger_mask)
Logs the given information to cout and to the database as a configure record (OrcConfigRecordP in Objectivity). The information should agree with that for the configure record, otherwise a warning is issued (but the information is still logged).
LOG_BEGIN (runnumber, partition_number, runtype, partition_name)
Logs the given information to cout and to the database (OrcRunRecordP in Objectivity).
LOG_END (runnumber, partition_number, runtype, partition_name)
Logs the given information to cout and to the database (OrcRunRecordP in Objectivity).
LOG_END (runnumber, pnumber, runtype, pname)
Logs the given information to cout
(OrcRunRecordP in Objectivity).
The information should agree with that for the corresponding BEGIN, otherwise
an warning is issued (the information is still logged).
SHUTDOWN
Terminates unix process.
Dataflow Proxy
Process name = dataflow
SMI Object name = <partition>::DATAFLOW
Provides interface for all dataflow interaction for a given partition. This process contains the actual odfPartition, odfPlatform and odfManager objects for the given partition.
States and Actions:
state action
STARTINGUP
SHUTDOWN
HALTED
RESET
SHUTDOWN
STANDBY
ALLOCATE (crate_mask)
SHUTDOWN
ALLOCATED
MAP
DEALLOCATE
REBOOT
SHUTDOWN
UPDATE_CRATE_MASK (crate_mask)
PARTITIONED
CONFIGURE (environment, triggers, runtype)
UNMAP
REBOOT
SHUTDOWN
CONFIGURED
BEGIN (runnumber)
UNCONFIGURE
REBOOT
SHUTDOWN
SYNCH
RUNNING
END_RUNNING
PAUSE
REBOOT
SHUTDOWN
PAUSED
RESUME
REBOOT
SHUTDOWN
REBOOTING
GOTO_STANDBY
SHUTDOWN
ERROR_CRATES (failed to partition requested crates)
RESET
ERROR_MAP
REBOOT
SHUTDOWN
RESET
ERROR_CONFIGINFO (failed to find/load the required calCycle file)
RESET
REBOOT
SHUTDOWN
ERROR_TIMEOUT (transition timed out)
RESET
ERROR_ (damage occurred on this transition)
REBOOT
SHUTDOWN
NOT_THERE
Action details:
RESET
Sets the state of the proxy to the previous state. This is usually sent
by run control after a transition timed out and the proxy is returned to
the state prior to the attempt. The transition can then be retried etc.
REBOOT
Reboots the crates in the allocated partition, if there is one.
SHUTDOWN
Dissolves the dataflow partition and exits.
MAP
Sends map(0) to dataflow partition.
UPDATE_CRATE_MASK (crate_mask)
Change the default crate mask, taken from database, to the specified one.
CONFIGURE (environment, triggers, runtype)
If environment=0, reads the file Orc/runtype.dat to determine the correct configuration key for the given run type. If environment!=0 then this is used as the configuration key. The cofiguration key is then used to look up the appropriate CalCycleTC file
(OrcXXXXXXXX.dat, where XXXXXXXXX is the hex. key value). If the CalCycleTC file is opened successfully, the the configure transition, configure(key,triggers) is sent tot the dataflow partition. A trigger mask of 0 defines the ensuing runs as
calibration runs. The CalCyleTC file contains the required set of datflow transitions which must be driven after a BEGIN command (see below).
UNMAP
Sends unmap(0) to dataflow partition.
BEGIN (runnumber)
Resets the current position in the CalCycleTC hierarchy to the top and sends begin(runnumber) to the dataflow partition. Completed dataflow transitions now automatically trigger further ones as specified in the CalCycleTC file specified on the
CONFIGURE transition. In a calibration sequence (trigger this will continue until the calibration is complete and the dataflow partition has been cycled all the way through to CONFIGURED again. For non calibration runs the dataflow will generally be
cycled through to ENABLED. Receipt of a DISABLE (from an END_RUNNING command will then initiate sequencing out to CONFIGURED. All the dataflow states within the RUNNING state (meta, major, minor and enable cycles) are hidden from Run Control.
UNCONFIGURE
Send command unconfigure(0) to the dataflow partition.
END_RUNNING
Sets an internal flag so that all outstanding calibration cycles will be skipped and the partition will step normally out to CONFIGURED. If in a non calibration run a diable is generated with disable(0) taking run.
PAUSE
Sends pause(0) to dataflow partition.
RESUME
Sends resume(0) to dataflow partition.
ALLOCATE (crate_mask)
The dataflow partition and manager objects are instantiated and the
dataflow partition is created, using the given crate mask.
DEALLOCATE
Dataflow partition is released, freeing the dataflow crates.
Detector Controls Proxy
Process name = <various>
SMI Object name = <various - follows ODC naming conventions>
Several detector control proxies exist, one for each detector controls "component". Each proxy has its own Domain and can be allocated to (controlled by) up to 1 run control partition. The detector controls proxies run independently of the Run Control,
providing monitoring and logging services for the detector controls. When allocated to a Run Control these proxies provide the interface through which the detector component may be synchronised (e.g. downloading of set points etc.)
States and Actions:
state action
ERROR
CLEAR
READY
CONFIGURE (key)
BEGIN_RUN (key)
CONFIG_LOCKED
END_RUN (key)
NOT_THERE
Action details:
CONFIGURE (key)
Initiates downloading of all confguration/setpoint data for the proxy concerned. Any required data/values are retrieved from the database using the given key.
BEGIN_RUN (key)
Warns the proxy that a run is beginning. The key is used to check that the detector component has been correctly configured. By default this will simply be a check that this key is the same as that in the most recent CONFIGURE transition. Configuration
data may now not be modified until the run has ended.
END_RUN (key)
Denotes end of run and unlocks the configuration data in the proxy so that it can be modified. The key is used as a consistency check with the most recent BEGIN_RUN.
OEPL3 Proxy
Process name = <OepML3ScriptProxy>
SMI Object name = <OEPL3>
Interface to OEP system. For details check with OEP group.
Logbook Proxy
Process name = <logbook>
SMI Object name = <LOGBOOK>
Interfaces to the Oracle electronic log book to log information at the
beginning and end of runs. If this proxy is absent, the run control will
continue as normal (but nothing will get logged into oracle).
state action
ERROR
RESET
READY
LOG_BEGIN_RUN (run_number, crate_mask, key, trigger_mask)
LOG_END_RUN (run_number)
Action Details
RESET(key)
Forces proxy state to READY.
LOG_BEGIN_RUN (run_number, crate_mask, key, trigger_mask)
Logs to specified information to oracle along with other start of run info.
determined independently by the proxy itself.
LOG_END_RUN (run_number)
Logs to specified information to oracle along with other start of run info.
determined independently by the proxy itself (e.g. events recorded).
Marker Proxy
Process name = <marker>
SMI Object name = <MARKER>
This proxy watches the output (XTC) file size and the run duration.
If either the file size or the run length exceed predetermined limits
(currently 20GB and 3 hours) then this proxy requests that the current
run be stopped and a new one started (called inserting a marker).
state action
ERROR
RESET
READY
START (runnumber)
RUNNING
STOP
RESET
MARKER_REQUEST
MARKER_SENT
Action Details
RESET(key)
START
Start the proxy monitoring the time and file size for the current run.
When a marker (stop/start run) is requested, the proxy goes to state MARKER_REQUEST.
STOP
Tells the proxy to stop monitoring, until told to START again.
MARKER_SENT
Informs the proxy that the marker was sent successfully. Currently the proxy
simply goes directly to RUNNING.
Monitor Proxy
Process name = <monitor>
SMI Object name = <MONITOR>
This proxy interfaces to the Odf dead time monitoring; principally so that
dead time can accurately be counted during the live time of the run.
state action
READY
START (crate_mask)
RUNNING
STOP
START (crate_mask)
CHANGE_UPDATE_RATE (period)
ERROR
RESET
Action Details
RESET(key)
Clear error and reset proxy to READY
START
Start the odf dead time monitoring with a predetermined update frequency.
CHANGE_UPDATE_RATE (period)
Changes the update frquency of the odf deadtime monitoring to <period> seconds.
STOP
Stop the current odf dead time monitoring.
Messenger Proxy
Process name = <messenger>
SMI Object name = <MESSENGER>
This proxy controls the audio alarm voice synthesizer and allows run control
to send voice alarms. The proxy manages its own list of messages, read in
from a file and specified by numbers. All communication with run control uses the message numbers.
state action
READY
MESSAGE (message)
REPEAT_MESSAGE
Action Details
MESSAGE(message)
Send the predefined message number to the voice sythesizer.
REPEAT_MESSAGE
Guess !
ShellCmd Proxy
Process name = <shellCmd>
SMI Object name = <SHELLCMD>
Provides a limited interface between the run control and the local
operating system, allowing execution of some OS commands.
state action
READY
EXECUTE (command, parameter)
Action Details
EXECUTE (command, parameter)
Execute the specified command with the given arguments.
Code_archiver Proxy
Process name = <codeArchiver>
SMI Object name = <CODEARCHIVER>
Used to archive a snapshot of all the binary and package versions currently being
used in the online system.
state action
ERROR
RESET
READY
ARCHIVE (run_number)
Action Details
ARCHIVE (run_number)
Archive the contents of the production release to a tar file; Run<runnumber>.tar.gz.
|
