running LO with a local non-embedded database under Qubes OS

Hi,

I have been having problems trying to get Firebird running as a database on the local machine, pending it arriving as an embedded db (currently targetted for LO 6.2).

Some of my problems have been due to using the Qubes OS, which is designed to be “the most secure” version of Linux (and endorsed as such by Snowden). I am guessing this is an issue with file permissions or locations. Qubes does clever stuff to make sure system changes roll back at reboot, whereas user changes are retained. It does this by running several virtual machines, known as “Qubes”, within a Xen hypervisor.

When using FlameRobin to create a user or a database on the same “Qube”, the FlameRobin settings get written to the current user’s home directory, which is preserved by Qubes when you next use the system. However firebird (as installed from the repos) makes changes to one or more files in the system space, which Qubes then rolls back thinking it is an unauthorised system change.

The consequence is that an inconsistency develops which seems difficult to resolve.

I am looking into this further, but would appreciate any suggestions of anyone else has alreadt found a workable solution to this. Please post an answer or a comment, as you feel appropriate

My current thought is to create a standalone “Qube” to act as a db server within the same physical machine and to connect internally via an internal IP address as if it were a remote database. That way Qubes understands what is going on better. (Or I understand better how to tell Qubes what is going on…)

I will post a howto if/when I get this working.

Duplicate question. This is already discussed under → connect base to firebird(server) - how??. Also, topic relates to “Qubes” issue and not LibreOffice.

hi Ratslinger

The question relates to the interaction of LO, FlameRobin, and the Qubes OS and is therefore of possible interest to anyone else trying to get firebird running as a split db on a qubes system.

Precisely becasue the Qubes aspects of the problem is going to be of little interest to most LO users, I have split off the Qubes specific stuff here, and removed it from the question you linked to.

I hope you think this is sensible…