-
Bug
-
Resolution: Unresolved
-
Low
-
None
-
6.3.1
-
None
-
9
-
Severity 2 - Major
-
0
-
Summary
When using the import feature of Bamboo with a non-trivial export file (many changesets, relevant changesets and test results) the import process is very slow if the database type used is MS SQL Server.
Environment
Bamboo 6.3.1
Database MS SQL server with the Microsoft driver
Steps to Reproduce
- Install a fresh Bamboo 6.3.1 instance with MS SQL server.
- During the installation import an exported file that has many changesets (in the order of 100000),
relevant changesets (in the order of 100000), commits (in the order of 100000), and test results (in the order of millions)
Expected Results
The import completes in a reasonable time (in the order of tens of minutes).
By comparison when using Postgres 9.4 as the external database type the import of the same export file completes in ~45 min.
Actual Results
The import takes an unreasonable amount of time to complete (more than 13 days).
Notes
Using a different database connector such as JTDS does not change the behavior.
Workaround
There is no known workaround.
[BAM-19726] Bamboo import using a non-trivial export file is very slow when the external database type is MS SQL Server
We face also the same issue, especially with file VariableDefinition.xML, this process need to be drastically improved
There are known performance problems with importing large export files in Bamboo. We strongly recommend using native DB migration tools for purpose of backup/restore (https://confluence.atlassian.com/bamboo/exporting-data-for-backup-289277255.html).
I believe this issue should be highlighted on this page .
This is so very disappointing that Atlassian doesn't inform users using MS SQL Server, about this known issue. How do you guys expect a user to think of this issue or reach here ?