fix(prover):fix prover panic infinite retry (#824)

Co-authored-by: georgehao <georgehao@users.noreply.github.com>

prover/client/client.go

@@ -165,7 +165,7 @@ func (c *CoordinatorClient) GetTask(ctx context.Context, req *GetTaskRequest) (*
 }
 
 // SubmitProof sends a request to the coordinator to submit proof.
-func (c *CoordinatorClient) SubmitProof(ctx context.Context, req *SubmitProofRequest) error {
+func (c *CoordinatorClient) SubmitProof(ctx context.Context, req *SubmitProofRequest) (needDelete bool, err error) {
 	var result SubmitProofResponse
 
 	resp, err := c.client.R().
@@ -175,27 +175,24 @@ func (c *CoordinatorClient) SubmitProof(ctx context.Context, req *SubmitProofReq
 		Post("/coordinator/v1/submit_proof")
 
 	if err != nil {
-		log.Error("submit proof request failed: %v", err)
-		return fmt.Errorf("submit proof request failed: %w", ConnectErr)
+		return false, fmt.Errorf("submit proof request failed: %v", err)
 	}
 
 	if resp.StatusCode() != 200 {
-		log.Error("failed to submit proof, status code: %v", resp.StatusCode())
-		return fmt.Errorf("failed to submit proof, status code not 200: %w", ConnectErr)
+		return false, fmt.Errorf("failed to submit proof, status code: %v", resp.StatusCode())
 	}
 
 	if result.ErrCode == types.ErrJWTTokenExpired {
 		log.Info("JWT expired, attempting to re-login")
 		if err := c.Login(ctx); err != nil {
-			log.Error("JWT expired, re-login failed: %v", err)
-			return fmt.Errorf("JWT expired, re-login failed: %w", ConnectErr)
+			return false, fmt.Errorf("JWT expired, re-login failed: %v", err)
 		}
 		log.Info("re-login success")
 		return c.SubmitProof(ctx, req)
 	}
 	if result.ErrCode != types.Success {
-		return fmt.Errorf("error code: %v, error message: %v", result.ErrCode, result.ErrMsg)
+		return true, fmt.Errorf("error code: %v, error message: %v", result.ErrCode, result.ErrMsg)
 	}
 
-	return nil
+	return true, nil
 }

prover/client/types.go

@@ -1,13 +1,9 @@
 package client
 
 import (
-	"errors"
-
 	"scroll-tech/common/types/message"
 )
 
-var ConnectErr = errors.New("connect coordinator error")
-
 // ChallengeResponse defines the response structure for random API
 type ChallengeResponse struct {
 	ErrCode int    `json:"errcode"`

prover/prover.go

@@ -162,7 +162,7 @@ func (r *Prover) proveAndSubmit() error {
 		proofMsg, err = r.prove(task)
 		if err != nil { // handling error from prove
 			log.Error("failed to prove task", "task_type", task.Task.Type, "task-id", task.Task.ID, "err", err)
-			return r.submitErr(task, true, message.ProofFailureNoPanic, err)
+			return r.submitErr(task, message.ProofFailureNoPanic, err)
 		}
 		return r.submitProof(proofMsg)
 	}
@@ -321,12 +321,14 @@ func (r *Prover) submitProof(msg *message.ProofDetail) error {
 	}
 
 	// send the submit request
-	if err := r.coordinatorClient.SubmitProof(r.ctx, req); err != nil {
-		if !errors.Is(errors.Unwrap(err), client.ConnectErr) {
-			if deleteErr := r.stack.Delete(msg.ID); deleteErr != nil {
-				log.Error("prover stack pop failed", "task_type", msg.Type, "task_id", msg.ID, "err", deleteErr)
-			}
+	needDeleteTask, err := r.coordinatorClient.SubmitProof(r.ctx, req)
+	if needDeleteTask {
+		if deleteErr := r.stack.Delete(msg.ID); deleteErr != nil {
+			log.Error("prover stack pop failed", "task_type", msg.Type, "task_id", msg.ID, "err", deleteErr)
 		}
+	}
+
+	if err != nil {
 		return fmt.Errorf("error submitting proof: %v", err)
 	}
 
@@ -335,7 +337,7 @@ func (r *Prover) submitProof(msg *message.ProofDetail) error {
 	return nil
 }
 
-func (r *Prover) submitErr(task *store.ProvingTask, isRetry bool, proofFailureType message.ProofFailureType, err error) error {
+func (r *Prover) submitErr(task *store.ProvingTask, proofFailureType message.ProofFailureType, err error) error {
 	// prepare the submit request
 	req := &client.SubmitProofRequest{
 		TaskID:      task.Task.ID,
@@ -347,12 +349,14 @@ func (r *Prover) submitErr(task *store.ProvingTask, isRetry bool, proofFailureTy
 	}
 
 	// send the submit request
-	if submitErr := r.coordinatorClient.SubmitProof(r.ctx, req); submitErr != nil {
-		if !errors.Is(errors.Unwrap(err), client.ConnectErr) {
-			if deleteErr := r.stack.Delete(task.Task.ID); deleteErr != nil {
-				log.Error("prover stack pop failed", "task_type", task.Task.Type, "task_id", task.Task.ID, "err", deleteErr)
-			}
+	needDeleteTask, submitErr := r.coordinatorClient.SubmitProof(r.ctx, req)
+	if needDeleteTask {
+		if deleteErr := r.stack.Delete(task.Task.ID); deleteErr != nil {
+			log.Error("prover stack pop failed", "task_type", task.Task.Type, "task_id", task.Task.ID, "err", deleteErr)
 		}
+	}
+
+	if submitErr != nil {
 		return fmt.Errorf("error submitting proof: %v", submitErr)
 	}