We should probably keep here a list of known issues and maybe other stuff.
The official release is available here
Installation Instruction
- Read and follow the instructions at SystemSetup
- Create a setup script that sets up ROOT, DataFlow?, OnlineSoftware, Java etc following the exampleSetup.[c]sh example
- Copy the file $SCT_DAQ_ROOT/installed/config/databases/Hardware_template.data.xml to $SCT_DAQ_ROOT/installed/config/databases/Hardware.data.xml and modify appropriately (replace the strings "localhost", "sbcHost", "AnalysisHost" with the appropriate machines and check that the binary tag is set correctly
- Run it!
If you have a real ROD, you will need to set it up by creating a configuration file similar to the ones in $SCT_DAQ_ROOT/installed/config/rods. You choose which file by setting the environment variable $SCTDAQ_ROD_CONFIGURATION_PATH
You may also want to change/edit the partition file. Do this with the $TDAQ_DATA_PATH variable. The examples included are:
- SCT_test.data.xml: This runs the test machine all on the localhost
- SCT_nodcs.data.xml: This runs the real system, but does not include the DCS partition
- SCT.data.xml: Includes the DCS partition
Version Information
Here is the output of SctRodDaqVersion -l which prints out much information about the compiled version:VersionString:SctRodDaq_2_00 $ MajorVersion: 2 Minor Version: 0 Modifier: 00 $ Version: 2000 BuildSystem: Linux pcdk 2.4.18-27.7.cb2 #6 Mon Apr 14 10:50:39 BST 2003 i686 unknown BuildHost: pcdk.hep.phy.cam.ac.uk BuildDate: 19:22:16 17 Nov 2003 RodDaq Version: SctRodDaq_2_00 CmtConfig: i686-rh73-gcc32 Dataflow Version: DF-00-05-00 OnlineVersion : online-00-19-01 RootVersion: 3.05/07 GccVersion: 3.2.3
Fixes in version 2.01
- Archiving Service doesn't crash/cause problems any more
- Noise occupancy works
Known Issues
Known issues with this release should be listed here.- There is a DSP bug that means that all the channels are read-out when charge is being injected. This means that any scan which injects charge actually returns (scan + 3*0 fC response). For this reason we have truncated threshold scans at 90mV. However, it still cause huge numbers of defects and problems for many tests.
- Problem with SctApi communicating with DSP code. The method awaitResponse sometimes times out, causing the equivalent of a hang.
- The modules need to have a hard reset after the pipeline test [ What's the symptom here? I thought this was just the configuration masks being wrong, which should be fixed by a subsequent load (ie initialiseAll). A hard reset is different... BJG] I think that we had problems getting data from chips afterwards, but that could have been the masks I suppose....not sure... Matt
- The full bypass test has some issues (we're working on this)
- Due to lack of PMG synchronisation, sometimes not all the processes are ready when the system is in the "Loaded" state. This lasts for only a few seconds.
- The archiving is still very slow when there are lots of defects.
- The version information is incorrect in the Java code
- Remove $ from end of version in C++ code
- There is a bug in the timewalk test (fixed in 2.1)