Don't crash with mulitple tasks of the same name

[johnml1135]

Fix #606.

---

This change is 

[codecov-commenter]

Codecov Report

Attention: Patch coverage is 0% with 2 lines in your changes missing coverage. Please review.

Project coverage is 64.50%. Comparing base (62091bb) to head (fbdae7f).

Files with missing lines	Patch %	Lines
...l.Machine.Shared/Services/ClearMLMonitorService.cs	0.00%	2 Missing ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##             main     #607      +/-   ##
==========================================
- Coverage   64.50%   64.50%   -0.01%     
==========================================
  Files         280      280              
  Lines       14112    14113       +1     
  Branches     1824     1824              
==========================================
  Hits         9103     9103              
- Misses       4348     4349       +1     
  Partials      661      661              

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

[Enkidu93]

What situation is this covering? When would there be two tasks with the same name? Aren't the names the engine or build ids or something?

Reviewed 1 of 1 files at r1, all commit messages.
Reviewable status: complete! all files reviewed, all discussions resolved (waiting on @johnml1135)

[ddaspit]

Reviewed 1 of 1 files at r1, all commit messages.
Reviewable status: all files reviewed, 1 unresolved discussion (waiting on @johnml1135)

---

src/Machine/src/Serval.Machine.Shared/Services/ClearMLMonitorService.cs line 77 at r1 (raw file):

                    .OrderBy(t => t.Created)
                    .Select((t, i) => (Position: i, Task: t))
                    .GroupBy(e => e.Task.Name)

The task name is the build id. Build ids are GUIDs. How are we getting duplicate build ids?

[johnml1135]

The failure (which has actually happened 30+ times) is multiple tasks names "Inferencev3". It causes a failure of the update script.

[johnml1135]

src/Machine/src/Serval.Machine.Shared/Services/ClearMLMonitorService.cs line 77 at r1 (raw file):

Previously, ddaspit (Damien Daspit) wrote…

The task name is the build id. Build ids are GUIDs. How are we getting duplicate build ids?

The name is different - the name is user-defined, while there is a separate internal ID assigned to each task.

[johnml1135]

src/Machine/src/Serval.Machine.Shared/Services/ClearMLMonitorService.cs line 77 at r1 (raw file):

Previously, johnml1135 (John Lambert) wrote…

The name is different - the name is user-defined, while there is a separate internal ID assigned to each task.

The name corresponds to our ID.

[ddaspit]

Reviewable status: all files reviewed, 1 unresolved discussion (waiting on @johnml1135)

---

src/Machine/src/Serval.Machine.Shared/Services/ClearMLMonitorService.cs line 77 at r1 (raw file):

Previously, johnml1135 (John Lambert) wrote…

The name corresponds to our ID.

Somehow non-Serval tasks end up in the list of tasks. First, we get all tasks that correspond to currently running builds. Then, we get all tasks that are currently in the Serval queues. Are these queries retrieving non-Serval tasks? Do others use the Serval queues?

[johnml1135]

src/Machine/src/Serval.Machine.Shared/Services/ClearMLMonitorService.cs line 77 at r1 (raw file):

Previously, ddaspit (Damien Daspit) wrote…

Somehow non-Serval tasks end up in the list of tasks. First, we get all tasks that correspond to currently running builds. Then, we get all tasks that are currently in the Serval queues. Are these queries retrieving non-Serval tasks? Do others use the Serval queues?

Mark put some jobs on production - I don't know why - I will ask him. That is the source of this issue.

[johnml1135]

src/Machine/src/Serval.Machine.Shared/Services/ClearMLMonitorService.cs line 77 at r1 (raw file):

Previously, johnml1135 (John Lambert) wrote…

Mark put some jobs on production - I don't know why - I will ask him. That is the source of this issue.

I talked with Mark - he is going to limit the number of simultaneous jobs to 2 (to not bomb our GPUs), but they may still have the same name. Let's put this fix in.

[ddaspit]

Reviewable status: complete! all files reviewed, all discussions resolved (waiting on @johnml1135)

[Enkidu93]

Reviewable status: complete! all files reviewed, all discussions resolved (waiting on @johnml1135)