Re: [Tracker] support for other database backends?

[Jussi Laako <jussi laako linux intel com>]

On 29.5.2012 14:17, Philip Van Hoof wrote:
> The reason for the N9 was that a long-hold press of the power button on
> the N9 results in a (at least for Tracker's process) uninterruptable and
> more importantly uninterceptable power shutdown of the device.
>
> If this were not the case it indeed wouldn't have been needed to ever do
> a fsync, as all shutdowns would do unmount prior to powerdown.

I think fsync is wrong way to do it and may still cause data corruption 
on flash due to large erase blocks that don't match filesystem sector 
sizes (unless you use some direct-flash filesystem like LogFS/UBIFS).

So log-structured filesystems are better approach, giving possibility to 
rewind back to last known working snapshot without ruining performance.

Sync writes with SQLite can cause really severe write stalls due to 
triggered filesystem write barriers.