Details
-
Enabler
-
Should have
-
Data Processing
-
-
-
4.5
-
3
-
0
-
Team_PSS
-
Sprint 5
-
-
-
- Tango controls training has been completed.
- New diagrams representing how PSS.LMC CTRL tango devices trigger cheetah pipelines have been produced for both Mid and Low.
- We have worked with the CREAM team to update the PSS.LMC CTRL devices. We are now able to, from a K8s deployment of PSS.LMC, control a real cheetah pipeline (as opposed to a simulated cheetah) on a physical server.
- We have also updated the tango device properties so that CTRL can control different pipeline types (e.g, single-pulse search/pulsar search), different data sources (files, streams) and cheetah logging levels.
- Additionally we updated the scheduling block -> cheetah config functionality so that we can produce an XML config that conforms to cheetah's requirements.
- We can show that CTRL can start cheetah listening for a UDP stream from a single beam, and then stop the pipeline.
- A document showing the steps is added below.
- Extra helm charts have been produced for those wanting to test this functionality without access to the physical server.
- Future work has been agreed between FO and the team.
Tango controls training has been completed. New diagrams representing how PSS.LMC CTRL tango devices trigger cheetah pipelines have been produced for both Mid and Low. We have worked with the CREAM team to update the PSS.LMC CTRL devices. We are now able to, from a K8s deployment of PSS.LMC, control a real cheetah pipeline (as opposed to a simulated cheetah) on a physical server. We have also updated the tango device properties so that CTRL can control different pipeline types (e.g, single-pulse search/pulsar search), different data sources (files, streams) and cheetah logging levels. Additionally we updated the scheduling block -> cheetah config functionality so that we can produce an XML config that conforms to cheetah's requirements. We can show that CTRL can start cheetah listening for a UDP stream from a single beam, and then stop the pipeline. A document showing the steps is added below. Extra helm charts have been produced for those wanting to test this functionality without access to the physical server. Future work has been agreed between FO and the team. SP-4248 Outcomes.pdf -
24.1
-
Stories Completed, Solution Intent Updated, Outcomes Reviewed, Satisfies Acceptance Criteria, Accepted by FO
-
-
AA2
Description
https://miro.com/app/board/uXjVKZ900uo=/?moveToWidget=3458764587830073316&cot=14
Beneficiaries (Who)
- PSI engineers; PSS developers
Benefit Hypothesis (Why)
The implementation and integration of PSS LMC will enable PSI and AIV engineers to control and configure the PSS pipeline using CSP.LMC or TMC, which will improve user experience of using the PSS software during testing e.g via using notebooks for the control and configuration commands. It will also align the method of controlling PSS with the that used for rest of the SKA software, thus integrating PSS into the SKA pipeline.
Description (What?)
This work will enable developers of PSS to contribute to the development of PSS LMC and to gain hands-on experience with the existing TANGO codebase, such that the amount of effort required in follow-up PIs can be estimated more accurately. As part of this process, the amount of effort to support PSS developers in the development of PSS LMC will be gauged by engaging with OMC teams and discussing the requirements of the PSS LMC interface with CSP.LMC and TMC. In addition, a standalone test demonstrating the basic functionality of the existing PSS CTRL TANGO device will be configured, using mock input and output configuration strings (i.e. the input will be manually sent to PSS CTRL and the output will not configure any pipeline), and the response of the device will be monitored and recorded.
Attachments
Issue Links
- mentioned in
-
Page Loading...