Fisheye + Crucible + Crowd : database usage

[Oscar Picasso]

Hi,

We plan to install Fisheye to view and manage git repositories, and use also Crucible for code review.

We will use Crowd for single sign on. Crowd will mostly authenticate user with our corporate LDAP but we will have some local crowd users.

For each of the product, what is the role of the database:

- fisheye. indexing? other? if the database crashes can we regenerate it from the git repositories (reindexing maybe?)

- crucible : same question as for fisheye

- crowd: storing local crowd users? other?

[Alexander Taler · Accepted solution]

FishEye and Crucible share a database, so only one backup.

Crucible stores all of your code reviews including comments and feedback. This is important historical information, particularly if you have any audit requirements for performing code reviews. There are also a number of user configuration settings.

FishEye includes an administrative option for scheduling backups to disk, all you need to do is copy it off or configure a remote drive. Doing backups is definitely the easier solution in this case.

[Chii]

Fisheye's index is not stored in the database, nor are the repository names and its configuration. In Fisheye, the database is used for storing user information (such as login, watches, and display preferences such as the diff options when viewing a diff).

Crucible uses the same user information as Fisheye, and also stores its review comments and other metadata in the database.

If you simply backup the database, it will mostly work, except that reviews with uploaded patches and/or attachments will be missing (those are stored on disk).

Backing up both the database, and home directory (set as FISHEYE_INST) will be the most comprehensive. You can also take advantage of the built-in backup feature, which will backup everything required to recreate the instance.

[V]

In general for any Atlassian product it is advisable to take back up of both Database and the Home directory.

[SimonS]

Please disregard the downvote - I accidentally clicked it and I can't undo it :( I'll replace your karma.

[Oscar Picasso]

Thanks for your answers.

If I use the built-in backup feature, does it also take care of backing up the git repositories?

If not, I guess I need to backup those git repositories separately. But in such a case there is a risk of desynchronization between the fisheye backup and the git backups. What is the best practice deal with that ?

[Oscar Picasso]

Actually, the fisheye documentation seems to provide the answers. Sorry for the noise.