The topology is dual stack, the JAVA UME exist in the ABAP stack. However during the downtime, the ABAP and the JAVA stack function on different instance numbers.
During the upgrade, SUM creates a temporary destination called SAPJup in the previous phase CONNECT-TEMP-ABAP. (Log file: <Path to SUM/sdt/log/SWITCH/CONNECT-TEMP-ABAP_xx.LOG) to allow the JAVA to connect to the ABAP UME.
- July 20, 2017 4:02:33 PM [Info]: Original RFC destination settings obtained successfully from the configuration manager.
- July 20, 2017 4:02:33 PM [Info]: Saving the temporary upgrade RFC destination settings under the following configuration path destinations/RFC/SAPJup.
- July 20, 2017 4:02:33 PM [Info]: The upgrade RFC destination settings were saved successfully under the following configuration path: SAPJup.
- July 20, 2017 4:02:33 PM [Info]: The UME service was updated successfully with the temporary upgrade RFC destination location.
- July 20, 2017 4:02:33 PM [Info]: The upgrade RFC destination name was successfully saved in the upgrade persistency.
The configuration of the SAPJup has similar entries: (To view the configuration, open configtool -> switch to the configuration editor mode -> configuration -> Destinations -> RFC -> Property Sheet SAPJup)
The configuration displays the SUM uses the jco.client.user as DDIC and the jco.client.client as the productive client and the jco.client.ahost and jco.client.sysnr is related to the host and system number where the ABAP is now functioning in the execution phase.
The error occurs when the password for DDIC in the productive client is separate from the password for DDIC in 000 client.
During the phase of configuration, the password for DDIC user for 000 clients is provided to the SUM tool as shown in the image below:
In the phase START-AS-JAVA, SUM tries to initiate the JAVA UME in ABAP by connecting to the productive client with DDIC user and the password for DDIC for 000 client.
As the passwords are different, the user id is not authenticated and ultimately gets locked because of incorrect login attempts and the java ume service fails to start, thus leading the phase to fail.
Solution
For overcoming the error, the password for the DDIC user in the productive client should be altered in the ABAP stack for matching the password for the DDIC in 000 client.
Follow the steps below:
1. Unlock ABAP system for allowing login to productive client (as listed by the property jco.client.client in the property sheet SAPJup) in ABAP using the OS command: tp unlocksys <SID> pf=<Path to SUM>/abap/var/DEFAULT.TPP
2. Login to ABAP system (running on hostname <jco.client.ahost> with the shadow system number <jco.client.sysnr>) in the productive client with a user who has access to SU01 transaction code.
3. Modify the password of DDIC user in the productive client for matching the password for DDIC for 000 client (and also unlock the DDIC user if its locked already because of incorrect login attempts).
4. Log out and login as DDIC in a productive client in ABAP with the new password (same password for DDIC for 000 client) to verify if the login is now successful.
5. Lock the system again by utilizing the command : tp locksys <SID> pf=<path to SUM>/abap/var/DEFAULT.TPP
6. Stop the JAVA instance manually, by utilizing the command: sapcontrol -nr <XX> -function Stop
7. Repeating the phase in SUM