fix: improve DRIFTED_FILES parsing and reduce file descriptor issues

- Add comprehensive debug output to track parsing steps - Fix DRIFTED_FILES extraction from Ansible output - Always output DRIFTED_FILES line (even when empty) for reliable parsing - Add ANSIBLE_CALLBACKS_ENABLED='' to prevent inotify exhaustion - Add KEEP_PLAYBOOK_LOG option for debugging - Add validation warning when OUT_OF_SYNC but no files found - Create DEBUGGING_GITOPS_STATUS.md guide
2026-04-22 23:03:34 +03:00 · 2026-04-22 23:03:34 +03:00 · 46b0bb449e
commit 46b0bb449e
parent 57911b7f52
1 changed files with 187 additions and 0 deletions
--- a/DEBUGGING_GITOPS_STATUS.md
+++ b/DEBUGGING_GITOPS_STATUS.md
@ -0,0 +1,187 @@
+# Debugging GitOps Status Issues
+
+This guide helps troubleshoot issues with the GitOps status reporting system.
+
+## Common Issue: Status shows OUT_OF_SYNC after deployment
+
+### Symptoms
+- You pushed changes to the repo
+- The deploy step succeeded
+- But `update-gitops-status` shows OUT_OF_SYNC
+- Changed files are not displayed
+
+### Root Causes
+
+#### 1. **Deployment didn't actually succeed**
+The deploy step might have failed silently or the configuration wasn't applied correctly.
+
+**How to check:**
+- Look at the deploy step logs in Woodpecker
+- SSH to the server and verify files match Git:
+  ```bash
+  diff /etc/rsyslog.conf files/rsyslog.conf
+  diff -r /etc/rsyslog.d/ files/rsyslog.d/
+  ```
+
+#### 2. **Parsing issue with DRIFTED_FILES output**
+The script might not be correctly extracting the file list from Ansible's output.
+
+**How to debug:**
+Run the status update script locally with debug mode:
+```bash
+export KEEP_PLAYBOOK_LOG=true
+./update-gitops-status.sh
+```
+
+This will save the playbook output to `drift-check-output.log`. Check:
+- Does the log contain `DRIFTED_FILES=` line?
+- What does the line look like exactly?
+- Are there ANSI color codes interfering?
+
+Look for these debug lines in the output:
+```
+DEBUG: Searching for DRIFTED_FILES in playbook output...
+DEBUG: Found DRIFTED_FILES pattern
+DEBUG: Raw line: ...
+DEBUG: Extracted value: '...'
+```
+
+#### 3. **Too many open files error**
+If you see "failed to create fsnotify watcher: too many open files":
+
+**Fixed in latest version:**
+- `.woodpecker.yml` now sets `ANSIBLE_CALLBACKS_ENABLED=""` and `ANSIBLE_GATHERING=explicit`
+- `update-gitops-status.sh` uses `ANSIBLE_CALLBACKS_ENABLED=""` when running playbooks
+- These settings prevent Ansible from exhausting inotify watches
+
+**If issue persists:**
+- The container might need privileged mode to adjust kernel parameters
+- Or reduce Ansible parallelism in inventory settings
+
+#### 4. **Ansible output format changed**
+If Ansible version changed, the debug output format might be different.
+
+**How to fix:**
+Check `drift-check-output.log` and adjust the parsing in `update-gitops-status.sh`:
+```bash
+# Current parsing (line ~110 in update-gitops-status.sh):
+DRIFTED_FILES_STR=$(echo "$DRIFTED_FILES_STR" | sed 's/.*DRIFTED_FILES=//' | sed 's/\x1b\[[0-9;]*m//g' | sed 's/".*$//' | xargs)
+```
+
+You might need to adjust the `sed` commands based on the actual format.
+
+## Testing the Fix
+
+### 1. Test locally
+```bash
+# Set up SSH key
+export SSH_PRIVATE_KEY="$(cat ~/.ssh/id_rsa)"
+
+# Run the script with debug output
+export KEEP_PLAYBOOK_LOG=true
+./update-gitops-status.sh
+
+# Check the log
+cat drift-check-output.log | grep -A 2 "DRIFTED_FILES="
+```
+
+### 2. Test in Woodpecker
+Push a small change and monitor the `update-gitops-status` step:
+```bash
+# Make a small comment change
+echo "# Test change $(date)" >> files/rsyslog.conf
+
+# Commit and push
+git add files/rsyslog.conf
+git commit -m "test: verify gitops status detection"
+git push
+
+# Watch the pipeline in Woodpecker UI
+# The update-gitops-status step should:
+# 1. Run deploy (apply.yml)
+# 2. Run drift-check immediately after
+# 3. Show SYNCED (because deploy just ran)
+# 4. Show no drifted files
+```
+
+### 3. Test drift detection (manual change on server)
+```bash
+# SSH to the server
+ssh rsyslog-lab
+
+# Make a manual change
+echo "# Manual change" >> /etc/rsyslog.conf
+
+# Wait for the cron job to run (every 2 minutes)
+# Or manually trigger it in Woodpecker
+
+# The status should now show:
+# - Status: OUT_OF_SYNC
+# - Files: rsyslog.conf
+```
+
+## Expected Behavior
+
+### After successful deployment (push to master)
+```
+Step 2/4: Analyzing drift detection results...
+  ✓ Status: SYNCED - server configuration matches Git
+  Total drift count: 0
+
+Step 3/4: Building JSON payload...
+  Generated JSON:
+  {
+    "repo": "rsyslog",
+    "server": "rsyslog-lab",
+    "sync_status": "SYNCED",
+    "drift_count": 0,
+    "files": [],
+    "last_check": "2026-04-22T14:30:00Z"
+  }
+```
+
+### When drift is detected (cron job or manual server change)
+```
+Step 2/4: Analyzing drift detection results...
+  ✗ Status: OUT OF SYNC - configuration drift detected
+    - Drift detected in: rsyslog.conf
+  Total drift count: 1
+
+Step 3/4: Building JSON payload...
+  Generated JSON:
+  {
+    "repo": "rsyslog",
+    "server": "rsyslog-lab",
+    "sync_status": "OUT_OF_SYNC",
+    "drift_count": 1,
+    "files": [
+      {"name": "rsyslog.conf"}
+    ],
+    "last_check": "2026-04-22T14:32:00Z"
+  }
+```
+
+## Quick Reference
+
+### Enable debug mode
+```bash
+export KEEP_PLAYBOOK_LOG=true
+./update-gitops-status.sh
+cat drift-check-output.log
+```
+
+### Manually run drift-check
+```bash
+ansible-playbook -i ansible/inventory/hosts.yml ansible/playbooks/drift-check.yml
+```
+
+### Manually run deployment
+```bash
+ansible-playbook -i ansible/inventory/hosts.yml ansible/playbooks/apply.yml
+```
+
+### Check server state
+```bash
+ssh rsyslog-lab "md5sum /etc/rsyslog.conf /etc/rsyslog.d/*.conf"
+md5sum files/rsyslog.conf files/rsyslog.d/*.conf
+```