Problem with JCMA sending wrong queries to postgres Database

Michael Bohlender November 16, 2023

We are just in the stage ot test migrating selected products of our on premises Jira Software to a cloud instance of Jira Software. App Assesment went fine, as we will discontinue to use all our add-ons since they were mainly used in the contect of business projects for time tracking, contact data etc. So we will bring Jira "back to the roots" during the migration, therefore the selected migration.

However I am facing the problem with the migration assistant, that it cannot read projects from the local instance as it sends wrong SQL queries to the database. As usual the PostgreSQL Database is case-sensitive when query terms are placed in double quotes. As such, queries where the table project is used as 

SELECT xyz FROM public."Project"

fail with unknown table errors from the dabase leading to corresponding SQL exceptions to be found in the atlassian-jira.log

Did anybody out there face the same problems and if yes, is there any solution / workaround that does NOT mean to switch the database system e.g. to MS-SQL just in order to prepare the migration?

Any help highly welcome 

Regards, Michael

2 answers

0 votes
Michael Bohlender December 3, 2023

So finally and fully frustrating, Atlassian managed to get it messed up to that point, that today after upgrading the JCMA to 1.10.9 it cannot be enabled anymore at all. And guess what: it fails to enable due to a metrics query

 select count(distinct "PROJECT"."ID") from "public"."Project"

which once again uses the completely wrong spelling for the table name when running on a PostgreSQL database.

In the meantime, I have even tried an empty Jira test setup on a second PGSQL server to see, if somewhere in the time since 2012 when our Jira was installed for the first time, things have changed that much in the database. But to my surprise tables have been created with exactly the same names as we see them on our production database.

Based on all the problems I have with this migration tool, we are currently even thinking about leaving Jira and Conluence behind and switch over to something else. A step we even did not think about in the worst night mares.

So I sincerely hope that anybody out there at Atlassian is willing and able to support us in this!

Regards, Michael

0 votes
Marc - Devoteam
Rising Star
Rising Star
Rising Stars are recognized for providing high-quality answers to other users. Rising Stars receive a certificate of achievement and are on the path to becoming Community Leaders.
November 16, 2023

Hi @Michael Bohlender 

You might be running into this issue, see KB article here

Michael Bohlender November 16, 2023

Hi @Marc - Devoteam ,

thanks for the hint but unfortunately not the cause. The schema tag is already present in the dbconfig, and besides some timeout and pool size values my file looks exactly as the one in the Kb article.

This is why I do not understand that the JCMA produces the erronous queries.

Marc - Devoteam
Rising Star
Rising Star
Rising Stars are recognized for providing high-quality answers to other users. Rising Stars receive a certificate of achievement and are on the path to becoming Community Leaders.
November 17, 2023

Hi @Michael Bohlender 

My advise would be to reach out to Atlassian Support.

they have access to Cloud instance backend, they can check the logging.

Like Michael Bohlender likes this
Michael Bohlender November 17, 2023

Hi @Marc - Devoteam 

error is happening on our local instance. I contacted our local Atlassian partner already for support.

Suggest an answer

Log in or Sign up to answer