SFTP Proxy Service Fails With The JMS Session Pool Is Shutting Down while trying to migrate back to orignal server after an automatic service migration (ASM) (Doc ID 2934317.1)

Last updated on MARCH 13, 2023

Applies to:

Symptoms

On : 12.2.1.4.0 version, OSB Build

ACTUAL BEHAVIOR
---------------
On SOA 12.2.1.4 SFTP Transport issue when trying to move back migratable targets to their original targets

Customer is testing high availability for that is using 2 node server (same can be seen on larger clusters) where they have 2 servers (1 and 2)

Each server is on different machines on different data centers

Under OSB cluster , server 1 is configured to migrate to server 2 (Auto-Migrate Excactly-Once Services) and vice versa

When server1 is shut down this migrates OK to server2, SFTP Transport keeps working with no issues

When server1 is bring back up customer is serve runs OK. Service still on server2, all services (SFTP transport) are working OK

When tries to migrate back to original server (on this case from server2 to server1) using weblogic console Environment / Clusters / osb cluster / Migratable Targets and then select server1 / control / check box on server1 and then "migrate button", then if OSB server 1 is pulling files at that moment starts failing with the error:

This is intermittent and seems to happen only when the service is SFTP is pulling files at the moment of the manual migration.

also all services using SFTP transport stop working, the only way to work around the issue is to restart the server1 (or the server where the errors are seen)

EXPECTED BEHAVIOR
-----------------------
That the migration is restored to the original server without disrupting the SFTP

STEPS
-----------------------
The issue can be reproduced at will with the following steps:
S1 is osb Server 1 and S2 is OSB Serverr2
1. On a 2 node cluster has osb running normally and configured cluster migration services from S1 to S2 and S2 to S1
2. stop the S1, jms modules are automatically migrated to S2, SFTP keeps working
3. Start S1, services remain on S2 SFTP keeps working as expected
4. In order to bring back high availablity (as services still on S2) manually migrate back form WLS console fro S2 to S1 for S1
5. if SFTP is processing files at that time the connection pool gets shut down and then all SFTP stops
6. The issue can be resolved by restarting S1

BUSINESS IMPACT
-----------------------
The issue has the following business impact:
Customer needs this to be highly available thus requesting a solution on how to be able to migrate services back once service is restored

Changes

Cause

My Oracle Support provides customers with access to over a million knowledge articles and a vibrant support community of peers and Oracle experts.