Missing columns after migration JIRA 6.3.15 to JIRA 7.4.5

Question

Hello,

I have trouble when upgration from JIRA 6.3.15 to JIRA 7.4.5
I use a clone of the databse of JIRA 6, to connect JIRA 7

For 1 server, everything is workking well, but for 2 other servers, i have error 500 with the following message when trying to access to a project :

java.lang.RuntimeException: java.lang.reflect.UndeclaredThrowableException
...
Caused by: com.atlassian.activeobjects.internal.ActiveObjectsInitException: bundle [com.pyxis.greenhopper.jira]
 at com.atlassian.activeobjects.osgi.TenantAwareActiveObjects$1$1$1.call(TenantAwareActiveObjects.java:95) [?:?]
 at com.atlassian.activeobjects.osgi.TenantAwareActiveObjects$1$1$1.call(TenantAwareActiveObjects.java:86) [?:?]
 at com.atlassian.sal.core.executor.ThreadLocalDelegateCallable.call(ThreadLocalDelegateCallable.java:38) [?:?]
 at java.util.concurrent.FutureTask.run(FutureTask.java:266) [?:1.8.0_102]
 at java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1142) [?:1.8.0_102]
 at java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:617) [?:1.8.0_102]
 ... 1 more
Caused by: java.lang.IllegalStateException: Could not find column 'LONG_QUERY' in previously parsed query!
 at net.java.ao.schema.helper.DatabaseMetaDataReaderImpl.getFields(DatabaseMetaDataReaderImpl.java:118) [?:?]
 at net.java.ao.schema.ddl.SchemaReader.readFields(SchemaReader.java:126) [?:?]
 at net.java.ao.schema.ddl.SchemaReader.readTable(SchemaReader.java:110) [?:?]
 at net.java.ao.schema.ddl.SchemaReader.access$000(SchemaReader.java:62) [?:?]
 at net.java.ao.schema.ddl.SchemaReader$1.apply(SchemaReader.java:99) [?:?]
 at net.java.ao.schema.ddl.SchemaReader$1.apply(SchemaReader.java:97) [?:?]
...

I had no errors in the UPGRADEHISTORY and UPGRADETASKHISTORY tables (all tasks to complete)

Comparing with a fresh JIRA 7.4.5 install, i noticed column LONG_QUERY is a nullable in AO_60DB71_SWIMLANE table (Jira Agile), and i other missing columns / tables :
LONG_QUERY in AO_60DB71_QUICKFILTER
LONG_QUERY in AO_60DB71_SUBQUERY
KAN_PLAN_ENABLED in AO_60DB71_RAPIDVIEW
GOAL in AO_60DB71_SPRINT
APPROVAL_STATE in AO_E8B6CC_ORGANIZATION_MAPPING (dvcs connector)
UPDATE_LINK_AUTHORISED in AO_E8B6CC_REPOSITORY_MAPPING
PRINCIPAL_ID in AO_E8B6CC_ORGANIZATION_MAPPING
Missing table AO_E8B6CC_ORG_TO_PROJECT
Missing table AO_E8B6CC_REPO_TO_PROJECT

So all missing columns / tables are Jira's tables. I don't understand why upgrading tasks are ok, but these columns/tables aren't created (others are) :
for example :

2018-11-21 11:46:32,752 JIRA-Bootstrap WARN      [o.o.c.entity.jdbc.DatabaseUtil] Field "description" of entity "AuditLog" is missing its corresponding column "DESCRIPTION"
2018-11-21 11:46:32,765 JIRA-Bootstrap INFO      [o.o.c.entity.jdbc.DatabaseUtil] Added column "DESCRIPTION" to table "audit_log"
2018-11-21 11:46:32,765 JIRA-Bootstrap WARN      [o.o.c.entity.jdbc.DatabaseUtil] Field "longDescription" of entity "AuditLog" is missing its corresponding column "LONG_DESCRIPTION"
2018-11-21 11:46:32,786 JIRA-Bootstrap INFO      [o.o.c.entity.jdbc.DatabaseUtil] Added column "LONG_DESCRIPTION" to table "audit_log"

I tried another DB clone ==> same results.
Did you experienced same problems ?

Thanks for your help,
Donatien

Andy Heinzer · Answer

Either way can work to upgrade Jira.   Personally I prefer to use the XML backup and restore that to a fresh install connected to a new empty database.  But when you do this, it's really important to make sure that you have restored the $JIRAHOME/data/ folder before you do this import.   I would not recommend copying/restoring the entire $JIRAHOME folder as it contains plugins that won't be compatible with Jira 7 that you had working with Jira 6.

More details in Upgrading Jira applications with a fallback method.  This method might take a bit longer to do vs the rapid upgrade method, but the ability to fallback and maintain your existing data is preserved in this method.

Donatien CARLOZ · Answer

Thanks for your answer Andrew.

I got this project few months ago, and the personn before me seems to successfully upgrade from JIRA6.3 to JIRA7.4.

I knew JIRA7.0 was strongly recommended, but i hoped there was a trick to avoid it :-)

So, i'll install first JIRA 7.0.

From your experience, do you recommend upgrading with xml file (restore JIRA6 xml file on JIRA7+JIRA7database), or cloning the database (so getting a JIRA 7 starting on a JIRA6 database) ?

Thanks,

Donatien

Andy Heinzer · Answer

You can't skip that many versions when trying to upgrade Jira.  All Jira 6.x versions should first be upgraded to a 7.0.x version before you can then upgrade to the latest versions.   This is detailed more in Skipping major versions when upgrading JIRA applications.

The 7.4.5 installation packages does not have all the upgradetasks needed to update all of your Jira/Jira Agile data.    By first upgrading to Jira Software 7.0.11 for example, you can update all your data here first.  I would also recommend that you update all the plugins that you use with Jira when you upgrade to this first, because that tends to be the only way to also update any plugin specific data that might in the Jira database.  This is especially important if you run Service Desk or plan to use Service Desk in the future.

Once that first upgrade is complete to 7.0.x version, you can then upgrade to your 7.4.5 or even a more recent version like 7.12.3