JIRA 6.3.4, migrate to other server, import mysql dump problem

[Hans Pesata]

Hi!

I am trying to perform a migration of our JIRA 6.3.4 installation on a LINUX machine to another LINUX machine.

By using mysqldump, I created a dump of the JIRA DB and I am importing the mysqldump file into the mysql JIRA DB on the destination machine.

I am copying the JIRA program directory and the JIRA home directory to the destination machine by using rsync using the --delete switch to exactly receive the source version of the file/directory structure.

Afterwards I am doing a reindex on the destination JIRA.

I realized, that the Base URL is still the original value but I suppose, that it should have been set to the value of the source machine ? Using mysql command line on the destination DB, I can see that the source machine BASE URL is present in the propertystring and changeitem tables, but not in the JIRA System GUI.

The issues of the source machine are visible, but when I try to update an issue I am receiving the following error:

"org.ofbiz.core.entity.GenericEntityException: while inserting: [GenericEntity:ChangeGroup][id,453410][author,sfe_admin][created,2015-03-18 15:15:08.288][issue,66821] (SQL Exception while executing the following:INSERT INTO changegroup (ID, issueid, AUTHOR, CREATED) VALUES (?, ?, ?, ?) (Duplicate entry '453410' for key 'PRIMARY'))
It seems that you have tried to perform an illegal workflow operation.
If you think this message is wrong, please contact your JIRA administrators."

Any help with what is going wrong or missing would be greatly appreciated!

Thanx in advance!

regards,
Hans

[Hans Pesata]

Hello!

Thank You for Your reply!

Why would the counter get messed up by doing a mysqldump export and an mysql import ?

Nevertheless I found this article which describes an operation to repair a situation like this:

https://confluence.atlassian.com/display/JIRAKB/Issue+Operations+are+Failed+by+the+PK+Violation+of+Database

And also why is the Base Url still the original one ? 

It is recommended to NOT use the XML backup but instead the native DB tools to do exports/imports but somehow this doesnt seem to work...

Any hints are very much appreciated!

Regards,
Hans 

[Timothy]

bq. Why would the counter get messed up by doing a mysqldump export and an mysql import ? No answer for that bq. And also why is the Base Url still the original one ? What you are doing is a copy and paste of the database. You will have to change the Base URL yourself.

[Hans Pesata]

The Base URL already shows the (old) destination address and not the one of the source machine which it should. Therefore I dont have to change it, like I have to when I use XML Backup. For me this implies, that something didnt get updated correctly, during the import... Regards, Hans

[Timothy]

I believe you've severely messed up the counter in your DB. 

Disclaimer & Note: I am not responsible for anything that happens after this

1. Stop your JIRA
2. Backup your DB
3. Navigate to your DB and the table called ChangeGroup and find the largest ID number
   1. You might also need to check the ChangeItem table which is referenced
4. Navigate to the SEQUENCE_VALUE_ITEM and check and if needed, increase the SEQ_ID for the SEQ_NAME of ChangeGroup to a large value than the one found in step 3
5. Restart JIRA and JIRA should pick up the latest higher number