File backup??

[Justin Jones]

Hello all, my first post here so be gentle... :)

We had a server failure(Ubuntu Linux) and i had to rebuild the server and re-install confluence, all that went pretty error free. Now when i set up confluence, i set it up with my lattest backup which was June 2013, id did a restore from /var/atlassian/application-data/confluence/restore that worked fine.

Now my question is... i still have accesss to the disk that has the old confluence files on it, and i woul like to have the most recent data from it, however its just the hard disk mounted on the new server in /mnt, i have no GUI access to it... how can i get the latest stuff off the old installation and restore to the new istallation?

Thanks..

Rob

Montreal, Canada

[Justin Jones]

OOoo... i think it worked... i will ask some people to check for thier entries they did prior to last Friday...

I just ran mysql -uUSER -pXXXX confluence <confluence_2013-08-30_06h35m.Friday.sql

Thanks !

Rob

Montreal, Canada

[Justin Jones]

Ok, I will try it and get back to you....

Thanks

Rob

Montreal, Canada

[Justin Jones]

Ok, so if i just import the Friday DB over teh June DB and re index it will be good?

As for editing now, no one has access to it now, until i resolve if i can bring it up to date more.

Thanks

Rob

Montreal, Canada

[Nic Brough -Adaptavist-]

Yes, if no-one has written to it since Friday (With any data you want to keep), then restoring the Friday database dump and reindexing should fix it.

You will need Confluence offline while that is done though!

[Nic Brough -Adaptavist-]

That could be quite difficult I'm afraid.

I assume you are talking mostly about the edits to Confluence pages that you have lost because you've effectively rolled the system back to June 2013.

These will still be in the database on the old disk, but that assumes the database was physically on that disk, and it wasn't damaged by the system failure. You'll need to find the database, read it, extract the updates and then the worst bit - apply them back to the current Confluence, merging them back in with other updates. That could be a really horrid piece of work.

To get you started on most of it, you need to start by getting access to the db - what type of database was it? (You probably need a real DBA to help you connect and extract here, not a part-timer like me)

[Justin Jones]

I have copy of the DB from Friday in teh form of an mySQL dump. SO i would need to add what was done from June 25th till Aug 30th.

What would i need to do, i do have some programmers here that are probably a bit better than me in DB commands and or processing/

Thanks for the quick reply!

Rob

Montreal, Canada

[Nic Brough -Adaptavist-]

Well, the copy is a good start, but that isn't the main worry.

The problem is that you need to replay over 30 days of data into the new database. My (very limited) understanding of a MySQL dump is that it is possible to convert it into the statements you need from it relatively easily, but you've got two problems with that

1) If anyone has edited your confluence between Aug 30th and now, their edits could well interfere with the ones being replayed

2) Finding the statements that were executed after June 25th and extracting them.

I'm really not a DBA and I'm not sure how easy finding and reconciling the transactions might be!

I'd be strongly tempted to do it the other way around. Backup the current confluence database and restore the data from Friday into it. That way, you'll only have to reconcile a weeks worth of changes.

(Oh and so I don't forget, the rebuild process is "stop confluence, run SQL, start confluence, make it re-index)