advLigoRTS issueshttps://git.ligo.org/cds/software/advligorts/-/issues2023-02-09T18:17:19Zhttps://git.ligo.org/cds/software/advligorts/-/issues/495adcparser.py parser needs improvement2023-02-09T18:17:19ZErik von Reisadcparser.py parser needs improvementIn particular, "INS" test fails on some channel names. In general, there is a better parser in the sequence checker. This ought to be reviewed and possibly reused for adcparser.py (or maybe do it all in one pass!).In particular, "INS" test fails on some channel names. In general, there is a better parser in the sequence checker. This ought to be reviewed and possibly reused for adcparser.py (or maybe do it all in one pass!).advligorts 5.2.0https://git.ligo.org/cds/software/advligorts/-/issues/494DTS1 Dolphin 4km Testing Issues2022-12-12T23:03:27ZEzekiel DohmenDTS1 Dolphin 4km Testing Issues# Needed Testing
# What resolved the last issue?
1. Switch reboots? <br>
a. Reboot with models running led to unstable IPCs <br>
b. Reboot with FEs off : TODO Result <br>
2. PX card link width settings? <br>
a. LR Test Seems t...# Needed Testing
# What resolved the last issue?
1. Switch reboots? <br>
a. Reboot with models running led to unstable IPCs <br>
b. Reboot with FEs off : TODO Result <br>
2. PX card link width settings? <br>
a. LR Test Seems to work in x8 or x 16 <br>
## Benchmark Dolphin Networks
Need to measure the timing variance, maybe the ~2us mean, ~4 us max is a sign of a switch issue.
# What are we doing that is causing the Dolphin switches to become unstable and need a reboot ?
## Dolphin API use?
TODO
## FE Restarts?
TODO
## Changing Dolphin Cards?
TODO
## Drivers/Firmware?
If the above tests don't cause the switches to go haywire, then we could conclude that the changing the firmware on the PX cards require a switch reboot.Ezekiel DohmenEzekiel Dohmenhttps://git.ligo.org/cds/software/advligorts/-/issues/492SDF save regular snap files2023-01-04T17:22:00ZErik von ReisSDF save regular snap filesWe'd also need some kind of pruning system to keep files manageable. The need for this isn't pressing since it seems the safe.snap settings for filter switches is good enough.We'd also need some kind of pruning system to keep files manageable. The need for this isn't pressing since it seems the safe.snap settings for filter switches is good enough.https://git.ligo.org/cds/software/advligorts/-/issues/490Standalone_edc does not have a way of reporting subscribed channels that are ...2023-02-09T18:20:12ZJonathan HanksStandalone_edc does not have a way of reporting subscribed channels that are in an error stateIf a EPICS CA subscription is in a state with a status != ECA_NORMAL there is no way to tell.
This was seen with the HWS ioc. On initialization it would only partially process its database, so you could caget a channel and get a value ...If a EPICS CA subscription is in a state with a status != ECA_NORMAL there is no way to tell.
This was seen with the HWS ioc. On initialization it would only partially process its database, so you could caget a channel and get a value (with an undefined status field) and the EDC would record it as 0.
It might be useful to track down issues like this to have the edc report what channels are currently giving status != ECA_NORMAL.
It should be noted that when status != ECA_NORMAL we do not have a value to collect, so our current behavior is correct as far as the data goes, this is just for diagnostics.advligorts 5.2.0Jonathan HanksJonathan Hankshttps://git.ligo.org/cds/software/advligorts/-/issues/489librts should allow re-getting the module2023-02-09T18:39:53ZErik von Reislibrts should allow re-getting the modulelibrts is particularly useful using jupyter notebooks, but the "load once" behavior of the models does not work well with a notebook. Cells in a notebook should be re-runnable without bad effect.
Ideally the model should be recreated, ...librts is particularly useful using jupyter notebooks, but the "load once" behavior of the models does not work well with a notebook. Cells in a notebook should be re-runnable without bad effect.
Ideally the model should be recreated, but barring that, the create_instance() method should return the existing model.
It would also be nice to have a way to reset filter historyEzekiel DohmenEzekiel Dohmenhttps://git.ligo.org/cds/software/advligorts/-/issues/485Switch cdsrfm to use different segments for long-range IPCS to each end station2022-10-31T17:24:22ZErik von ReisSwitch cdsrfm to use different segments for long-range IPCS to each end stationProtects against accidentally writing IPCs to the wrong end station if adapters are re-ordered.Protects against accidentally writing IPCs to the wrong end station if adapters are re-ordered.https://git.ligo.org/cds/software/advligorts/-/issues/483RCGTEST: Add mask check ot EPICS ctrl in2022-10-28T17:24:36ZErik von ReisRCGTEST: Add mask check ot EPICS ctrl inWritten procedure for the EPICS ctrl in part does not test that EPICs value changes are blocked when MASK input is 0. This check needs to be added if it doesn't exist, or brought into the report if it does exist.Written procedure for the EPICS ctrl in part does not test that EPICs value changes are blocked when MASK input is 0. This check needs to be added if it doesn't exist, or brought into the report if it does exist.https://git.ligo.org/cds/software/advligorts/-/issues/482Packaging: pin ldas-tools-framecpp to a specific version2022-10-28T12:11:43ZErik von ReisPackaging: pin ldas-tools-framecpp to a specific versionWe need to guarantee the right version and make it easy when promoting packages to see whether framecpp needs to be promoted.We need to guarantee the right version and make it easy when promoting packages to see whether framecpp needs to be promoted.https://git.ligo.org/cds/software/advligorts/-/issues/481AI chassis WD error should appear as a DAC error on the status line2022-10-27T19:00:37ZErik von ReisAI chassis WD error should appear as a DAC error on the status linehttps://git.ligo.org/cds/software/advligorts/-/issues/480Export daqd/local_dc/transport tests in packaging so that we can run the ctes...2022-11-02T19:10:26ZJonathan HanksExport daqd/local_dc/transport tests in packaging so that we can run the ctest based integration tests against the packaged bitsThe integration tests that are run via ctest could be adapted to work against /usr/bin and a local build directory allowing us to run the integration test outside of ctest (ie against built packages). This would aid in validation of the...The integration tests that are run via ctest could be adapted to work against /usr/bin and a local build directory allowing us to run the integration test outside of ctest (ie against built packages). This would aid in validation of the daqd and other components by testing the exact bits that get installed in a package.Jonathan HanksJonathan Hankshttps://git.ligo.org/cds/software/advligorts/-/issues/475RCG test: Some x2ats tests don't test what they claim.2022-12-23T19:13:24ZErik von ReisRCG test: Some x2ats tests don't test what they claim.daqLib Decimation Filter Test
tests some filter module loaded with coefficients, not the actual DAQ decimation in code. If the real decimation were to change, this test would never detect it. Could be a unit test, or a test on ATS w...daqLib Decimation Filter Test
tests some filter module loaded with coefficients, not the actual DAQ decimation in code. If the real decimation were to change, this test would never detect it. Could be a unit test, or a test on ATS with a special switch to decimate fast channels beyond what's needed for their rate.
This is RCG-0500-T in the RCG test procedure.
IOP decimation test (RCG-0501-T) has the same issue.https://git.ligo.org/cds/software/advligorts/-/issues/472Enable arbitrary decimation of ADC data2023-05-28T20:34:08ZErik von ReisEnable arbitrary decimation of ADC dataTo any power of two. Right now 8x decimation is not included, nor anything over 32x.
Here are some documents detailing the ADC decimation filters.
https://dcc.ligo.org/T1600066/
might have enough info to re-create the design of the f...To any power of two. Right now 8x decimation is not included, nor anything over 32x.
Here are some documents detailing the ADC decimation filters.
https://dcc.ligo.org/T1600066/
might have enough info to re-create the design of the filter.
https://dcc.ligo.org/T1500075/
also has some info.
We can either design the filter at model build, or assemble the filter from two equivalent filters, e.g. build an 8x filter from a 2x and 4x. It's not clear if the phase shift of combining filters is acceptable, but it's probably fine.https://git.ligo.org/cds/software/advligorts/-/issues/471Double precision for AWG2022-10-19T18:00:10ZJoseph BetzwieserDouble precision for AWGWe would like to request double precision AWG injection capability (and all related necessary adjustments to the RCG itself as well as AWG) to allow for injection of things like double precision DARM error on the test stand.We would like to request double precision AWG injection capability (and all related necessary adjustments to the RCG itself as well as AWG) to allow for injection of things like double precision DARM error on the test stand.https://git.ligo.org/cds/software/advligorts/-/issues/467awtpman failed on bootup x2iscey2022-10-13T21:57:52ZErik von Reisawtpman failed on bootup x2isceyFailed because of missing dependency. Maybe it tried to start too soon after the kmod was loaded?Failed because of missing dependency. Maybe it tried to start too soon after the kmod was loaded?https://git.ligo.org/cds/software/advligorts/-/issues/466offset not set properly in gpstime module with symetricom, possible race cond...2022-10-13T23:46:39ZJameson Rollinsjameson.rollins@ligo.orgoffset not set properly in gpstime module with symetricom, possible race condition in initialization?At the 40m the gpstime dkms modules are initializing with bad offsets. The calc_gps_offset.py script appears to be running, but the offset is not taken by the module:
```
Oct 7 16:00:42 c1ioo systemd-modules-load[191]: Inserted module ...At the 40m the gpstime dkms modules are initializing with bad offsets. The calc_gps_offset.py script appears to be running, but the offset is not taken by the module:
```
Oct 7 16:00:42 c1ioo systemd-modules-load[191]: Inserted module 'gpstime'
Oct 7 16:00:42 c1ioo kernel: [ 6.396375] gpstime: loading out-of-tree module taints kernel.
Oct 7 16:00:42 c1ioo kernel: [ 6.396881] Spectracom GPS card on bus 7; device 0
Oct 7 16:00:42 c1ioo kernel: [ 6.397039] gpstime: INFO - TSYNC PIC BASE 0 address = 0xfa9fe000
Oct 7 16:00:42 c1ioo kernel: [ 6.397157] gpstime: INFO - Remapped (____ptrval____)
Oct 7 16:00:42 c1ioo kernel: [ 6.699048] gpstime_sysfs_gps_offset_store success - new value = 0
Oct 7 16:00:42 c1ioo kernel: [ 6.742323] gpstime_sysfs_gps_offset_store success - new value = 0
Oct 7 16:00:42 c1ioo kernel: [ 11.406830] gpstime: INFO - TSYNC NOT receiving YEAR info, defaulting to by year patch
Oct 7 16:00:42 c1ioo kernel: [ 11.406992] gpstime: INFO - date = 280 days 23: 0:970873243
Oct 7 16:00:42 c1ioo kernel: [ 11.407093] gpstime: INFO - bcd time = 970873243 sec 309 milliseconds 367 microseconds 309362890 nanosec
Oct 7 16:00:42 c1ioo kernel: [ 11.407257] gpstime: INFO - Board sync = 1
Oct 7 16:00:42 c1ioo kernel: [ 11.417283] gpstime: INFO - TSYNC NOT receiving YEAR info, defaulting to by year patch
Oct 7 16:00:42 c1ioo kernel: [ 11.417444] gpstime: INFO - date = 280 days 23: 0:970873243
Oct 7 16:00:42 c1ioo kernel: [ 11.417546] gpstime: INFO - bcd time = 970873243 sec 319 milliseconds 820 microseconds 319815570 nanosec
Oct 7 16:00:42 c1ioo kernel: [ 11.417709] gpstime: INFO - Board sync = 1
```
Our suspicion is that there is maybe a race condition on initialization, where the script is being executed by udev before the module is fully initialized and able to accept the offset.Jonathan HanksJonathan Hankshttps://git.ligo.org/cds/software/advligorts/-/issues/465model build/run should have no dependence on IFO or any other arbitrary variable2022-10-17T15:49:55ZJameson Rollinsjameson.rollins@ligo.orgmodel build/run should have no dependence on IFO or any other arbitrary variableA common problem is that models configured for one IFO can't be run on another IFO. This is because we have insisted on the IFO and SITE variables being encoded in various places in the model build. This seem superfluous, and causes ma...A common problem is that models configured for one IFO can't be run on another IFO. This is because we have insisted on the IFO and SITE variables being encoded in various places in the model build. This seem superfluous, and causes many headaches. There should be nothing IFO specific about the models, and they should be portable between systems. The build system should eliminate any reference to IFO and SITE.https://git.ligo.org/cds/software/advligorts/-/issues/462RCGTEST: add sdf tests to ATS tests2022-12-23T18:01:56ZErik von ReisRCGTEST: add sdf tests to ATS testsSDF tests should be added to x2ats automated tests.
There's an older script here for automated testing that might be used as reference.
https://redoubt.ligo-wa.caltech.edu/viewvc/cds_user_apps/trunk/cds/test/scripts/python/sdf/SDF tests should be added to x2ats automated tests.
There's an older script here for automated testing that might be used as reference.
https://redoubt.ligo-wa.caltech.edu/viewvc/cds_user_apps/trunk/cds/test/scripts/python/sdf/automated test updatehttps://git.ligo.org/cds/software/advligorts/-/issues/460ATS test report needs be cleaned up2022-10-19T22:59:09ZErik von ReisATS test report needs be cleaned up* The test report must get version from code. Right now it's fixed at 4.2. Either that or just update it.
* Uses screen shots as examples, but these may be out of date. It might be useful to get live screenshots during the run. I do...* The test report must get version from code. Right now it's fixed at 4.2. Either that or just update it.
* Uses screen shots as examples, but these may be out of date. It might be useful to get live screenshots during the run. I don't think the old screenshots are helpful.
* Reference test number in document. These numbers don't exist yet but will be in the test procedure. Also, use consistent wording with the rest of the ATS system so that it's easy to tell what test you're looking at.
* If a test is not run for some reason, the report does not show this.
* A summary of pass/fail at the top would be good.
* A git summary would be good: git commit hash, any branch or tag names, last committer, last commit message.automated test updatehttps://git.ligo.org/cds/software/advligorts/-/issues/458Sequencer issues2022-10-11T00:20:43ZJonathan HanksSequencer issuesThe 5.0 sequencer has issues.
* cannot select the table view (table selection)
* cannot select a wildcard pattern (show all/substring)
* empty fec.snap (is this an issue)?The 5.0 sequencer has issues.
* cannot select the table view (table selection)
* cannot select a wildcard pattern (show all/substring)
* empty fec.snap (is this an issue)?Jonathan HanksJonathan Hankshttps://git.ligo.org/cds/software/advligorts/-/issues/456Bram slagmolen cannot get 2nd ADC to work with 5.0.12022-10-09T03:44:57ZErik von ReisBram slagmolen cannot get 2nd ADC to work with 5.0.1He is using 1pps signal with 65khz clock. Works in 5.0.0~dev0
Getting ADC timeout on last ADC on two different systems. If set the last to time_shift = 1, ADC0 fails. If both set to time_shift=1, ADC2 fails.He is using 1pps signal with 65khz clock. Works in 5.0.0~dev0
Getting ADC timeout on last ADC on two different systems. If set the last to time_shift = 1, ADC0 fails. If both set to time_shift=1, ADC2 fails.