htwkmooc/poseidon
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

Refactor Nomad Recovery

Browse Source 
from an approach that loaded the runners only once at the startup
to a method that will be repeated i.e. if the Nomad Event Stream connection interrupts.
This commit is contained in:
Maximilian Paß
2023-10-23 14:36:14 +02:00
committed by Sebastian Serth
parent b2898f9183
commit 6b69a2d732
22 changed files with 211 additions and 120 deletions
Download Patch File Download Diff File Expand all files Collapse all files

2
internal/api/ws/codeocean_writer.go
View File

@@ -107,7 +107,7 @@ func (cw *codeOceanOutputWriter) Close(info *runner.ExitInfo) {
case errors.Is(info.Err, context.DeadlineExceeded) || errors.Is(info.Err, runner.ErrorRunnerInactivityTimeout):
cw.send(&dto.WebSocketMessage{Type: dto.WebSocketMetaTimeout})
case errors.Is(info.Err, runner.ErrOOMKilled):
cw.send(&dto.WebSocketMessage{Type: dto.WebSocketOutputError, Data: runner.ErrOOMKilled.Error()})
cw.send(&dto.WebSocketMessage{Type: dto.WebSocketOutputError, Data: dto.ErrOOMKilled.Error()})
case errors.Is(info.Err, nomad.ErrorAllocationCompleted), errors.Is(info.Err, runner.ErrDestroyedByAPIRequest):
message := "the allocation stopped as expected"
log.WithContext(cw.ctx).WithError(info.Err).Trace(message)