---
title: "Workflow transitions fail after update"
canonical: "https://support.appfire.com/space/SUPPORT/3706519761/Workflow%20transitions%20fail%20after%20update"
format: markdown
---
# Summary

This article addresses issues where Comala Document Management (CDM) workflow transitions (e.g., Draft to QA Review) fail to complete and / or metadata is not rendered after an app update. This behavior is typically caused by one of two distinct scenarios: corrupted search index journal files or network connectivity restrictions preventing internal loopback calls.

---

## Scenario 1: Corrupted search index journal files

### Problem

Workflow transitions fail to update the page state in the UI, and metadata or issue numbers do not populate correctly. The logs show a `NumberFormatException` related to the Confluence journal.

### Cause

Whenever a workflow transition occurs, CDM requests Confluence to reindex the page content. If the Confluence search index journal tracker files (located in the local home directory) are empty or corrupted (containing a "null" ID), the indexing task cannot be enqueued. This causes the entire transaction to fail mid-execution.

The following error may be visible in the Confluence logs:

```plaintext
com.atlassian.cache.CacheException: org.springframework.dao.DataAccessResourceFailureException: Failed to read id for journal 'main_index': Cannot parse null string
Caused by: java.lang.NumberFormatException: Cannot parse null string
    at com.atlassian.confluence.impl.journal.FilesystemJournalStateStore.getMostRecentId(FilesystemJournalStateStore.java:44)
```

### Solution

1. **Stop Confluence** (on all nodes if running a Data Center cluster).
2. Navigate to the Confluence Local Home directory on the server: `<confluence-local-home>/journal/`.
3. Inspect the files in the `journal` folder. Look for any empty (0-byte) or corrupted files related to `main_index` or `edge_index`.
4. Delete or move these files out of the folder. Confluence will automatically regenerate clean journal trackers upon startup.
5. **Start Confluence**.
6. Navigate to **Administration** (cog icon) > **General Configuration** > **Content Indexing**.
7. Run a **Full search re-index**.
8. Navigate to **Administration** > **Manage apps** > **Comala Document Management** > **Modules** and ensure the **velocity-allowlist-cdx-apps** module is enabled.
9. Test the workflow transition on a **new page**.

---

## Scenario 2: Network connectivity and loopback restrictions

### Problem

Workflow transitions or support package exports run extremely slowly or time out. The system logs may show connection timeout warnings when the server attempts to reach its own Base URL.

### Cause

Comala Document Management and Confluence may require the ability to make internal loopback calls to the site's own Base URL. If outbound security group rules or firewalls restrict the server from talking to its own Load Balancer IPs (hairpin NAT), these internal calls will time out, causing workflow failures.

### Solution

1. **Check Logs for Timeouts:** Look for connection timeout warnings in the Confluence logs when attempting to reach the site's own Base URL (e.g., `https://confluence-test.example.org`).
2. **Update Security Groups/Firewalls:** Ensure that the Confluence nodes are permitted to communicate with the IP addresses associated with their own hostname/Load Balancer.
  - *Note:* In AWS environments, Load Balancer IPs change regularly. Ensure rules account for these changes or use a more dynamic security group configuration.
3. **Verify Loopback:** Ensure "hairpin NAT" or loopback traffic to the Base URL is allowed from the nodes.
4. **Reindex:** After restoring connectivity, it is recommended to reset the Confluence index and perform a full site re-index to ensure all pending updates are processed.

---

## Additional Resources

- [How to clear Confluence plugins cache](https://confluence.atlassian.com/confkb/how-to-clear-confluence-plugins-cache-297664846.html)