advLigoRTS issueshttps://git.ligo.org/cds/software/advligorts/-/issues2023-01-19T04:34:29Zhttps://git.ligo.org/cds/software/advligorts/-/issues/146RCG no longer producing caQtDM screens on front-end builds2023-01-19T04:34:29ZKeith ThorneRCG no longer producing caQtDM screens on front-end buildsIn earlier development this year, Rolf had created caQtDM (*.ui) versions of the front-end screens. These were used in the post-build scripts to create the same screens as for MEDM *.adl)
However in test builds on l1boot0 it is noted th...In earlier development this year, Rolf had created caQtDM (*.ui) versions of the front-end screens. These were used in the post-build scripts to create the same screens as for MEDM *.adl)
However in test builds on l1boot0 it is noted that it can find these. And if I check the disks on the large test stand, we see that the *.ui screens have not been built since May28
<pre>
[keith.thorne@x2portal x2lsc]$ ls -l *GDS_TP*
-rw-rw-r--. 1 controls advligorts 41141 Aug 7 17:06 X2LSC_GDS_TP.adl
-rw-rw-r--. 1 controls advligorts 222804 May 28 20:31 X2LSC_GDS_TP.ui
</pre>
I could not find which branch had the template files. git log did not seem to help4.0.0Erik von ReisErik von Reishttps://git.ligo.org/cds/software/advligorts/-/issues/144DAQ KB/S rates do not match between old daqd and producer2020-08-26T21:56:09ZKeith ThorneDAQ KB/S rates do not match between old daqd and producerWe now have the older data rates reported by the daqd processor, and newer producer-thread rates. But when comparing them (see [https://alog.ligo-la.caltech.edu/TST/index.php?callRep=13697](https://alog.ligo-la.caltech.edu/TST/index.php?...We now have the older data rates reported by the daqd processor, and newer producer-thread rates. But when comparing them (see [https://alog.ligo-la.caltech.edu/TST/index.php?callRep=13697](https://alog.ligo-la.caltech.edu/TST/index.php?callRep=13697)) they don't agree. Old FE KB/S of 53,618 and new PRDCR KB/S of 52,005. It appears one is in KiB/s and another in kB/s4.0.0Jonathan HanksJonathan Hankshttps://git.ligo.org/cds/software/advligorts/-/issues/143Add EPICS outputs to stand-alone EDCU2020-08-26T17:17:28ZKeith ThorneAdd EPICS outputs to stand-alone EDCUSee https://alog.ligo-la.caltech.edu/TST/index.php?callRep=13676
For better monitoring, the following channels would be useful
- GPS time (allows monitoring that the software is running)
- DAQ data rate (# KB/sec)
- Uptime (to monitor r...See https://alog.ligo-la.caltech.edu/TST/index.php?callRep=13676
For better monitoring, the following channels would be useful
- GPS time (allows monitoring that the software is running)
- DAQ data rate (# KB/sec)
- Uptime (to monitor restarts)4.0.0Jonathan HanksJonathan Hankshttps://git.ligo.org/cds/software/advligorts/-/issues/141EzCaRead.pm has mismatch in Conn/Err output2020-07-31T17:40:04ZKeith ThorneEzCaRead.pm has mismatch in Conn/Err outputin src/epics/util/lib/EzCaread.pm, the second EPICS variable has suffix _CONN in the header structure (line 22) but suffix _ERR in the fromExp section (Line 72). This leads to an error in build when output is not sent to a terminator
/...in src/epics/util/lib/EzCaread.pm, the second EPICS variable has suffix _CONN in the header structure (line 22) but suffix _ERR in the fromExp section (Line 72). This leads to an error in build when output is not sent to a terminator
/var/cache/advligorts/rcg-4.0.0~dev5/src/fe/x2tstezca/x2tstezca.c:52:50: error: ‘TST’ {aka ‘struct TST’} has no membe
r named ‘X2_EXT_TEST_AO_ERR’; did you mean ‘X2_EXT_TEST_AO’?
pLocalEpics->tst.AO_READ_CONN = pLocalEpics->tst.X2_EXT_TEST_AO_ERR;
^~~~~~~~~~~~~~~~~~
X2_EXT_TEST_AO4.0.0Keith ThorneKeith Thornehttps://git.ligo.org/cds/software/advligorts/-/issues/139The standalone_edc doesnt export the epics variables it claims to2020-07-23T22:54:37ZJonathan HanksThe standalone_edc doesnt export the epics variables it claims toThe connected channel count, not connected count, and total channels.The connected channel count, not connected count, and total channels.4.0.0Jonathan HanksJonathan Hankshttps://git.ligo.org/cds/software/advligorts/-/issues/137High dcu (ecat) models added to par2020-07-17T23:51:38ZErik von ReisHigh dcu (ecat) models added to parHigh dcu numbers should be excluded from .par files.
I think I already made my change, maybe I forgot to do an MR.High dcu numbers should be excluded from .par files.
I think I already made my change, maybe I forgot to do an MR.4.0.0Erik von ReisErik von Reishttps://git.ligo.org/cds/software/advligorts/-/issues/136The standalone edcu shouldn't need a .par file for local_dc to work with it.2020-07-16T17:55:11ZJonathan HanksThe standalone edcu shouldn't need a .par file for local_dc to work with it.local_dc gets rate/dcuid info from a .par file. This is ok for models, it is completely unneeded for the standalone edc.local_dc gets rate/dcuid info from a .par file. This is ok for models, it is completely unneeded for the standalone edc.4.0.0Jonathan HanksJonathan Hankshttps://git.ligo.org/cds/software/advligorts/-/issues/135The systemd unit for the standalone edc is incorrect, uses 'I' instead of 'l'2020-07-15T20:58:02ZJonathan HanksThe systemd unit for the standalone edc is incorrect, uses 'I' instead of 'l'The wrong argument value is used, cap I instead of lower l.The wrong argument value is used, cap I instead of lower l.4.0.0Jonathan HanksJonathan Hankshttps://git.ligo.org/cds/software/advligorts/-/issues/134rtcds build World stops at first error2020-07-31T17:40:04ZErik von Reisrtcds build World stops at first errorThis is a problem on initial build when almost every model produces errors from missing channels in the chan file.This is a problem on initial build when almost every model produces errors from missing channels in the chan file.4.0.0Erik von ReisErik von Reishttps://git.ligo.org/cds/software/advligorts/-/issues/133standalone_edc overflows the stack2020-07-15T07:22:57ZJonathan Hanksstandalone_edc overflows the stackThis is visible in a package build, not a plain from source build as the debian packaging adds -fstack-protection and -fstack-protection-strong (or something like that) to the build.
This was tracked down to the boost::lock_free queue a...This is visible in a package build, not a plain from source build as the debian packaging adds -fstack-protection and -fstack-protection-strong (or something like that) to the build.
This was tracked down to the boost::lock_free queue api, the spsc (single producer/single consumer) queue. This is used to pass diagnostic information between the main epics thread and the network server that can give a dump of all the channels that are not currently connected. Upon investigation it was found that the api is suprising, the pop call when given a pointer will copy all available entries to the pointer, instead of a single entry. This is good in that it can allow efficient clearing of the queue, but surprising if you are not expecting this.4.0.0Jonathan HanksJonathan Hankshttps://git.ligo.org/cds/software/advligorts/-/issues/129IOPs running on secondary timing in test stands aren't updating.2020-06-29T17:05:14ZErik von ReisIOPs running on secondary timing in test stands aren't updating.IOP models on front ends without IO Chassis aren't updating.
Dolphin is working. Recent changes to the how these models are configured as part of removal of timing master/slave terminology are suspect.IOP models on front ends without IO Chassis aren't updating.
Dolphin is working. Recent changes to the how these models are configured as part of removal of timing master/slave terminology are suspect.4.0.0Rolf BorkErik von ReisRolf Borkhttps://git.ligo.org/cds/software/advligorts/-/issues/128When a GDS broadcast list is specified and it contains channels that do not e...2020-06-18T21:02:41ZJonathan HanksWhen a GDS broadcast list is specified and it contains channels that do not exist the daqd should crashThis represents a time when config has going out of sync, we should abort with a useful error message. This requires manual intervention to fix and issues may be missed for a long time if we do not abort immediately when we see an issue.This represents a time when config has going out of sync, we should abort with a useful error message. This requires manual intervention to fix and issues may be missed for a long time if we do not abort immediately when we see an issue.4.0.0Jonathan HanksJonathan Hankshttps://git.ligo.org/cds/software/advligorts/-/issues/127Compile error building some models2020-06-29T17:04:40ZErik von ReisCompile error building some modelsOn x2seiey in DTS1, the big test stand at LLO:
`In file included from /opt/rtcds/rtbuild/rcg-4.0.0~dev13/src/fe/x2iopseiey/x2iopseiey.c:1634:
/usr/share/advligorts/src/src/fe/controllerIop.c: In function ‘fe_start_controller’:
/usr/shar...On x2seiey in DTS1, the big test stand at LLO:
`In file included from /opt/rtcds/rtbuild/rcg-4.0.0~dev13/src/fe/x2iopseiey/x2iopseiey.c:1634:
/usr/share/advligorts/src/src/fe/controllerIop.c: In function ‘fe_start_controller’:
/usr/share/advligorts/src/src/fe/controllerIop.c:624:27: error: too many arguments to function ‘iop_dac_write’
dkiTrip = iop_dac_write( expect_delays );
^~~~~~~~~~~~~
In file included from /usr/share/advligorts/src/src/fe/controllerIop.c:71,
from /opt/rtcds/rtbuild/rcg-4.0.0~dev13/src/fe/x2iopseiey/x2iopseiey.c:1634:
/opt/rtcds/rtbuild/rcg-4.0.0~dev13/src/fe/x2iopseiey/../../include/drv/no_ioc_timing.c:274:1: note: declared here
iop_dac_write( void )
^~~~~~~~~~~~~
`
Also affects other models.4.0.0Rolf BorkRolf Borkhttps://git.ligo.org/cds/software/advligorts/-/issues/125nort2020-06-22T20:24:04ZErik von ReisnortA regression of #111 and #114. No RT models are showing up in testpoint.par, where they should not be.
The offending command was rtcds install --allA regression of #111 and #114. No RT models are showing up in testpoint.par, where they should not be.
The offending command was rtcds install --all4.0.0Erik von ReisErik von Reishttps://git.ligo.org/cds/software/advligorts/-/issues/123Errors in awg output when run from systemd2020-06-11T02:07:25ZErik von ReisErrors in awg output when run from systemdRolf found that awg output, when tested through decimation filters, produced errors (bad magnitude, phase, coherence) from time to time, between awg output and filter output.
The errors do not occur when run as root from the command l...Rolf found that awg output, when tested through decimation filters, produced errors (bad magnitude, phase, coherence) from time to time, between awg output and filter output.
The errors do not occur when run as root from the command line.
Example good and bad results attached.
[baddecfilters.xml](/uploads/a83d9c2b6762ce032958080c7416fc76/baddecfilters.xml)
[Template_Dec_Filters.xml](/uploads/d6e3862d09c0e5388e6616c87ed09a43/Template_Dec_Filters.xml)4.0.0Erik von ReisErik von Reishttps://git.ligo.org/cds/software/advligorts/-/issues/122Update 'rtcds lsmod' script to support Dolphin IX, replacement for open-mx2020-07-31T17:40:04ZKeith ThorneUpdate 'rtcds lsmod' script to support Dolphin IX, replacement for open-mxThe rtcds lsmod currently checks for open-mx and Dolphin DX drivers.
This needs to be update the check for Dolphin IX drivers (dis_kosif, dis_ix_ntb, dis_ix_dma, dis_irm). We are no longer using the open-mx driver).
Should also consider...The rtcds lsmod currently checks for open-mx and Dolphin DX drivers.
This needs to be update the check for Dolphin IX drivers (dis_kosif, dis_ix_ntb, dis_ix_dma, dis_irm). We are no longer using the open-mx driver).
Should also consider add-on to verify that mx_stream replacements are running (local_dc, cps_xmit)4.0.0Erik von ReisErik von Reishttps://git.ligo.org/cds/software/advligorts/-/issues/121Convert the advligorts-transport-dix to use the new ligo-dolphin packages2020-07-06T19:52:46ZJonathan HanksConvert the advligorts-transport-dix to use the new ligo-dolphin packagesThis will allow the dkms package to be used and not tie daqd systems to a specific kernel.This will allow the dkms package to be used and not tie daqd systems to a specific kernel.4.0.0Jameson Rollinsjameson.rollins@ligo.orgJameson Rollinsjameson.rollins@ligo.orghttps://git.ligo.org/cds/software/advligorts/-/issues/120cps_xmit does not need the -d gds_tp_dir param.2020-06-08T22:22:05ZJonathan Hankscps_xmit does not need the -d gds_tp_dir param.A mixup in parameters caused a lot of work, until we realized we were passing -d delay instead of -D delay. It should have failed.A mixup in parameters caused a lot of work, until we realized we were passing -d delay instead of -D delay. It should have failed.4.0.0Jonathan HanksJonathan Hankshttps://git.ligo.org/cds/software/advligorts/-/issues/119The receiver in the pub_sub code does not properly detect late messages2020-06-08T22:22:05ZJonathan HanksThe receiver in the pub_sub code does not properly detect late messagesIn this context a late message is a message that arrived after the receiver was done accepting messages for that (gps, cycle) pair, but while there was still room to count it.
This is a dropped message. When we mark it as late we can t...In this context a late message is a message that arrived after the receiver was done accepting messages for that (gps, cycle) pair, but while there was still room to count it.
This is a dropped message. When we mark it as late we can tell if it could have potentially been used if a longer delay had been used.
The current system does not change the state flags, so messages are not marked late, causing broken metrics.4.0.0Jonathan HanksJonathan Hankshttps://git.ligo.org/cds/software/advligorts/-/issues/116cdsrfmmonitor.py calls python2 scripts2020-05-21T21:25:07ZErik von Reiscdsrfmmonitor.py calls python2 scriptscdsrfmmonitor.py fails when calling cdscfg/stdenv.py because stdenv.py is not compatible with python.
This should be fixed in cdscfg, but this problem breaks cdsrfm, so an issue should remain open here until fixed.cdsrfmmonitor.py fails when calling cdscfg/stdenv.py because stdenv.py is not compatible with python.
This should be fixed in cdscfg, but this problem breaks cdsrfm, so an issue should remain open here until fixed.4.0.0Erik von ReisErik von Reis