Tucson Test Stand¶

This section contains site specific variations for the Tucson test stand.

Resources¶

LOVE: http://love1.tu.lsst.org
LOVE (k8s): http://love.tu.lsst.org
Argo CD: https://tucson-teststand.lsst.codes/argo-cd
Chronograf: https://chronograf-tucson-teststand-efd.lsst.codes/
Nublado: https://tucson-teststand.lsst.codes/
Rancher: https://rancher.tu.lsst.org (1)
Slack: #rubinobs-tucson-teststand

(1) Need to get kubeconfig file from here. File a Jira ticket with Tucson IT for access. Once able to log into Rancher:

Select the pillan cluster.
Click the Kubeconfig File button in top-right.
Near bottom of dialog, click the download link.

Non-Production Systems¶

The Tucson test stand operates all CSCs and systems on the production domain.

Bare Metal Machines¶

LOVE: love1.tu.lsst.org
T&S CSCs: tel-hw1.tu.lsst.org
Kubernetes: Can be done from own machine, just need kubeconfig file and kubectl installed.
- Systems run on the pillan cluster.
- Can also use: https://k8slens.dev/.
ATCamera (Tony Johnson): auxtel-mcm.tu.lsst.org
CCCamera (Tony Johnson): comcam-mcm.tu.lsst.org
ATArchiver (Steve Pietrowicz): auxtel-archiver.tu.lsst.org
CCArchiver (Steve Pietrowicz): comcam-archiver.tu.lsst.org
Calibration Systems (Patrick Ingraham): loonie.tu.lsst.org

LOVE Summary View¶

The overall system summary state view is called Summary State.

Checking the Number of Federations¶

This uses a script in https://github.com/lsst-ts/k8s-admin. Run ./feds-check from a machine with kubectl and the proper kubeconfig file.

Shutdown DM and Camera Services¶

Shutdown/Cleanup daemon on Archiver machines:
- docker stop ospl-daemon
- docker rm ospl-daemon
Shutdown Camera OCS Bridges:
- ATCamera: sudo systemctl stop ats-ocs-bridge.service
- CCCamera: sudo systemctl stop comcam-ocs-bridge.service
Shutdown Camera Daemons
- sudo systemctl stop opensplice.service
- Command is the same everywhere.

Shutdown LOVE¶

This needs to be done from love1.

Uses the docker-compose-admin scripts in tucson-teststand/love1 directory.
- ./shutdown_love
- ./shutdown_daemon

Shutdown T&S Bare Metal Services¶

Handle tel-hw1:

Uses the docker-compose-admin scripts in tucson-teststand/tel-hw1 directory.
- ./shutdown_atmcs_atp
- ./shutdown_m1m3
- ./shutdown_daemon

Handle calibration systems:

Log into the machines listed in that section then stop and remove all running containers.

Interacting with Kubernetes¶

Commands can be executed from your own machine with kubectl and the proper kubeconfig file.

Shutdown Main Daemon¶

The main daemon on TTS runs on Kubernetes. Shut it down by deleting the deployment under the ospl-main-daemon app on Argo CD.

Update Configuration¶

Gather the branch for the configurations and version number for ts_ddsconfig.
Uses the docker-compose-admin scripts in tucson-teststand directory.
Directories to update:
- /deploy-lsstts/docker-compose-ops (love1, tel-hw1)
- /deploy-lsstts/ts_ddsconfig (love1, tel-hw1)
- /deploy-lsstts/LOVE-integration-tools (love1)
- sudo ./update_repo <repo path> <branch or version>
This will fail if the branch has local modifications. At that point you may as well just do the job manually. Here is one way to do that:
- cd /deploy-lsstts/<problem directory>
- git status
- sudo git reset –hard origin/<current ticket branch>
- Return to the docker-compose-admin scripts and run the update_repo command again.

Startup Main Daemon¶

The main daemon on TTS runs on Kubernetes and will be handled by the sync_apps.py script. This will be detailed in the next section

Startup Minimal Kubernetes System¶

This replaces most of step 6.3 in the main document. Follow the first three bullet points in that step and then continue the process with the next steps.

python sync_apps.py -p
csc-cluster-config, ospl-config and ospl-main-daemon apps will be synced automatically.
Once the ospl-main-daemon app is synced, the script will pause.
Check the logs on Argo CD UI to see if daemon is ready.
Type go and enter to move onto syncing the ospl-daemon app
Once the ospl-daemon app is synced, the script will pause.
Check the logs on Argo CD UI to see if daemons are ready.
Type go and enter to move onto syncing the kafka-producers app.
Script will again pause once the kafka-producers are synced.
The kafka-producers use a startup probe, so once all of the pods show a green heart, type go and enter to move onto syncing the love app.
Once the love app is synced, stop here and return to step 6.4 in the main document.
Make sure you leave the script running.

Startup LOVE¶

This needs to be done from love1.

Uses the docker-compose-admin scripts in tucson-teststand/love1 directory.
- ./launch_daemon
- Ensure daemon is ready before proceeding.
- ./launch_love

Startup T&S Bare Metal Services¶

Handle tel-hw1

Uses the docker-compose-admin scripts in tucson-teststand/tel-hw1 directory.
- ./launch_daemon
- Ensure daemon is ready before proceeding.
- ./launch_atmcs_atp
- ./launch_m1m3

Enabled CSCs¶

If proceeding with integration testing, the CSCs will be brought to ENABLED state as part of that process. All of the startup processes maybe necessary for recovering the TTS from any maintenance. In this case, all of the CSCs must be returned to ENABLED state. The following components will automatically transition to ENABLED state when launched:

Watcher
ScriptQueue:1
ScriptQueue:2
DSM:1
DSM:2

For the other components, leverage the following scripts. Required configurations will be given for each script execution.

Note

Both ATCamera and CCCamera must be in OFFLINE_AVAILABLE state before putting them into ENABLED state.

auxtel/enable_atcs.py

athexapod: ncsa
atdome: current
ataos: current

auxtel/enable_latiss.py

atcamera: Normal
atspectrograph: current

maintel/enable_mtcs.py

mtm1m3: Default
mthexapod_1: default
mthexapod_2: default

maintel/enable_comcam.py
```
cccamera: Normal
```

set_summary_state.py

data:
  -
    - DIMM:1
    - ENABLED
    - current
  -
    - DIMM:2
    - ENABLED
    - current
  -
    - WeatherStation:1
    - ENABLED
    - default