Issue
After Business Intelligence Exchange (BIX) runs in an affected environment, a subsequent data-flow leader node termination can cause data-flow leader election to fail. When this occurs, Queue Processors (QPs) might process records extremely slowly or stop processing altogether.
The issue does not occur immediately when BIX runs. It is triggered when the data-flow leader node is terminated after BIX has run.
Symptoms and Impact
Users might observe one or more of the following symptoms:
- One or more QPs process extremely slowly or stop processing.
- Functionality that depends on a QP stops working as expected.
- Newly created objects might not appear in search results because the pySASIncrementalIndexer QP is not processing records.
- Multiple unrelated QPs might experience processing problems at the same time.
The reported behavior might not appear to be related to BIX.
Affected releases
-
Pega Platform™ 25.1.4
-
Pega Platform 26.1.1
Steps to reproduce
- Use an environment running Pega Platform '25.1.4 or '26.1.1.
- Run BIX in the environment.
- Stop the data-flow leader node.
- Observe that one or more QPs process extremely slowly or stop processing.
Root Cause
Running BIX can create a condition in which a subsequent stop of the data-flow leader node prevents successful leader election. Without an elected leader, the affected data-flow processing cannot proceed normally, causing QPs to process records extremely slowly or stop processing.
Temporary mitigation
Restarting the cluster can temporarily restore normal processing. However, a cluster restart does not correct the underlying regression.
Solution
Apply the hotfix that corresponds to your Pega Platform release:
|
Pega Platform |
Hotfix |
|
25.1.4 |
HFIX-D4096 |
|
26.1.1 |
HFIX-D4092 |
The hotfixes revert the related change that causes the issue.
You can request this hotfix through the My Support Portal by selecting Existing Hotfix as the ticket type. For details on creating a support ticket, refer to the My Support Portal FAQs.