Move from M2 to M2a
When upgrading from M2 to M2a, can I just install M2a and copy our existing database (or do I need a database upgrade/conversion) ?
thx Daan. |
13 answers
When upgrading from M2 to M2a, can I just install M2a and copy our existing database (or do I need a database upgrade/conversion) ? As I couldn't find any information concerning migration from M2 to M2a (like it was with M1 and M2) I made a test with copping the folder 'SERVER_PATH/server/repositoryDB' from M2 to M2a. I ran RTC 0.6M2a, connected to the server and everything seemed to work fine (all the data was were it should be). If something doesn't work, it should be considered whether all the configuration files (e.g.teamserver.properties, tomcat-users.xml, etc.) were coppied as well. Of course it is recommended to make a backup of the old server, the repository data and configuration files in case something goes wrong. |
Thanks,
To be sure, I took the "long" way using reptools export/import. Daan. |
Hmm, looks like fulltext search is not working anymore.
Do I have to manualy recreate indexes ? regards Daan. |
Hmm, looks like fulltext search is not working anymore. It's a good point! I didn't notice this issue, but it appeared the same in my case. However in the server folder you will find folder named 'workitemindex', there is the index kept. I just copied this folder from the old version of the server to the new one and fulltext search started working again. Do you mean the same by saying that you wanted to 'manualy recreate indexes'? or did you mean something else? kind regards, Jakub |
Well, I thought that there might be some service running to create the fulltext serach indexes, which requires some manual override after an upgrade.
(a bit like the "locate" database on unix systems, which also needs to index the filesystem before you can actualy use it) regards Daan. |
There is allready a "workitemindex" directory in our migrated M2a jazz directory. From the looks of it, this was created after changes were made in our new (M2a) environment.
Not knowing the internals of the fulltext search feature, my guess would be that the index is only updated when you change/add something in the jazz-repository. In our case where we did not copy the "workitemindex" directory during the migration from M2 to M2a, only changes done after the migration are included in fulltext searches. A command/feature is needed which scans the complete repository and creates/fills the "workitemindex" from scratch. regards Daan. |
Dan,
Check Work Item 6504 https://jazz.net/jazz/web/projects/Jazz%20Project#action=com.ibm.team.workitem.viewWorkItem&id=6504 Remember you can always subscribe to WI that you are interested in. Elisabeth |
Not sure work-item 6504 is adresses this issue. The workitem concerns indexing on database level.
The "workitemindex" directory for the jazz fulltext-search feature looks like something outside the derby databse. regards Daan. |
There is allready a "workitemindex" directory in our The index format has changed during M2, that's why a new index directory is created. The migration tool basically takes care of recreating the index, so after an export/import, the full text search should work. If this was not the case, maybe there was something in the migration log? -- Patrick |
After migration from M2 to M2a (using export/import) there was no "workitemindex" directory.
During the export/import there were no errors in the log (output from repotools commandline) regards Daan. |
Your answer
Dashboards and work items are no longer publicly available, so some links may be invalid. We now provide similar information through other means. Learn more here.