Before reporting an issue

[X] I have read and understood the above terms for submitting issues, and I understand that my issue may be closed without action if I do not follow them.

Area

core

Describe the bug

We released keycloak 26 and functional and performance wise everything is good.

However around once a day a duplicate insert session happens , which results in 20.000 error messages in a 10 second time frame.

We have reduced this number by overwriting PersistentSessionsWorker, with a back-off time of 100ms instead of "0" (during 5 instead of 10) which is in the code:

Retry.executeWithBackoff((iteration) ->

, Duration.of(10L, ChronoUnit.SECONDS), 0);

The duplicate itself concerns 1 specific session id (1 of the 10.000 ?) around once a day, and doesn't seem to be a functional big issue:

HHH100501: Exception executing batch [java.sql.BatchUpdateException: Duplicate entry 'b9f91c53-ec37-4aaf-b54d-12f793c58d09-1' for key 'OFFLINE_USER_SESSION.PRIMARY'], SQL: insert into OFFLINE_USER_SESSION (BROKER_SESSION_ID,CREATED_ON,DATA,LAST_SESSION_REFRESH,REALM_ID,USER_ID,VERSION,OFFLINE_FLAG,USER_SESSION_ID) values (?,?,?,?,?,?,?,?,?)

(repeated 20.000 times of same entry id.)

Version

26.0.2

Regression

[ ] The issue is a regression

Expected behavior

when retry a insert session statement or others, use a backoff time > 0ms to avoid a bulk of same error messages.

Actual behavior

To many error messages because of back-off delay 0ms.

How to Reproduce?

generate a sql exception (like duplicate insert).

Anything else?

No response