Repository Sync Troubleshooting

# Check repository status
kubectl get repositories -n <namespace>

# Check repository conditions
kubectl describe repository <repo-name> -n <namespace>

# Verify sync configuration
kubectl get repository <repo-name> -n <namespace> -o yaml | grep -A5 sync

# Check repository synchronization logs
kubectl logs -n porch-system deployment/porch-controllers | grep "sync.*<repo-name>"

# Check for sync errors
kubectl logs -n porch-system deployment/porch-controllers | grep "<repo-name>.*error"

# Check secret exists (secret name is referenced in repository spec)
kubectl get repository <repo-name> -n <namespace> -o jsonpath='{.spec.git.secretRef.name}'
kubectl get secret <secret-name> -n <namespace>

# Verify secret data
kubectl get secret <secret-name> -n <namespace> -o yaml

# Recreate secret with correct credentials
kubectl delete secret <secret-name> -n <namespace>
kubectl create secret generic <secret-name> \
  --namespace=<namespace> \
  --type=kubernetes.io/basic-auth \
  --from-literal=username=<user> \
  --from-literal=password=<token>

# Porch controller will automatically retry authentication based on error-specific intervals
# See error handling documentation for retry timing details

# For immediate retry, re-register the repository with correct credentials:
kubectl delete repository <repo-name> -n <namespace>
kubectl delete secret <secret-name> -n <namespace>  # repo reg will create a new secret
porchctl repo reg <repo-url> --name <repo-name> --repo-basic-username <user> --repo-basic-password <token>

# Valid cron formats (5 fields: minute hour day month weekday)
porchctl repo reg <repo> --sync-schedule "*/10 * * * *"  # Every 10 minutes
porchctl repo reg <repo> --sync-schedule "0 */2 * * *"   # Every 2 hours
porchctl repo reg <repo> --sync-schedule "0 9 * * 1-5"   # 9 AM weekdays

# Invalid examples to avoid:
# "10 * * * *"     # Missing minute field
# "* * * * * *"    # Too many fields (6 instead of 5)

# Check Porch logs
kubectl logs -n porch-system deployment/porch-server

# Look for repository synchronization errors
kubectl logs -n porch-system deployment/porch-server | grep "repositorySync.*<repo-name>"

# Check repository accessibility
git ls-remote <repo-url>  # For Git repos

# Inspect the condition and the commit Porch expects to find in the history
kubectl describe repository <repo-name> -n <namespace>
kubectl get repository <repo-name> -n <namespace> -o jsonpath='{.status.gitCommitHash}'; echo

# Confirm whether the recorded commit is REACHABLE FROM the configured branch's
# history on the new remote. This must match what Porch validates: the commit must
# be an ancestor of (or equal to) the branch HEAD, not merely present somewhere in
# the repository. Use --single-branch so unrelated branches aren't fetched (a plain
# clone fetches every branch, so a bare existence check like `cat-file -e` can pass
# when the commit lives only on another branch — the very case Porch rejects).
git ls-remote <new-url> <branch>                              # current HEAD of the configured branch
git clone --branch <branch> --single-branch <new-url> /tmp/check
git -C /tmp/check merge-base --is-ancestor <recorded-hash> HEAD \
  && echo "reachable: sync will succeed" \
  || echo "NOT reachable (missing or on a different line of history): sync will be rejected"

# Option A: point at a URL whose branch history still contains the recorded commit
# (e.g. a true mirror/fork of the original, or the original itself moved forward)
kubectl patch repository <repo-name> -n <namespace> --type merge \
  -p '{"spec":{"git":{"repo":"<matching-url>"}}}'

# Option B: revert to the original URL to clear the error
kubectl patch repository <repo-name> -n <namespace> --type merge \
  -p '{"spec":{"git":{"repo":"<original-url>"}}}'

# Check runOnceAt field was set
kubectl get repository <repo-name> -n <namespace> -o jsonpath='{.spec.sync.runOnceAt}'

# Verify timestamp is in future
date -u  # Compare with runOnceAt value

# Check one-time synchronization logs
kubectl logs -n porch-system deployment/porch-controllers | grep "runOnceAt"

porchctl repo reg https://github.com/example/repo.git

kubectl get repositories -n <namespace>
kubectl get repositories --all-namespaces

porchctl repo sync <repo> --run-once 5m
porchctl repo sync <repo> --run-once "2024-12-01T15:00:00Z"

# Increase controller log level
kubectl patch deployment porch-controllers -n porch-system -p '{"spec":{"template":{"spec":{"containers":[{"name":"porch-controllers","args":["--v=2"]}]}}}}'

# Watch repository changes
kubectl get repositories -w -n <namespace>

# Monitor events
kubectl get events -n <namespace> --field-selector involvedObject.kind=Repository

# Repository sync logs
kubectl logs -n porch-system deployment/porch-controllers | grep "sync.*<repo-name>"

# Next sync time from status
kubectl get repository <repo-name> -n <namespace> -o jsonpath='{.status.nextFullSyncTime}'

# For Git repositories
git clone <repo-url>
find . -name "Kptfile" -type f  # Should find package directories

# Check branch exists
git branch -r | grep <branch-name>

--repositories.full-sync-frequency=1h
--repositories.health-check-frequency=5m

porchctl repo sync --all --all-namespaces