Re: Logging/RAID

"Axel Dörfler"  <[email protected]> wrote:
> "Ingo Weinhold" <bonefish-CFLBMwTPW48UNGrzBIF7/[email protected]> wrote:
> > > No, I meant:
> > > meta data = file and dir inodes, etc
> > > user data = actual file data
> > > 
> > > IIRC, only the former is journalled by BFS.
> > Oh, is that true?! Didn't know that.
> 
> It's exactly like that: logging file data would be ways to expensive. 
> Also, it wouldn't be doable with the current BFS log architecture (it 
> only has a maximum of 2048 blocks in the cache).
> Losing file data with BFS (and almost any other journaled file 
> system, 
> btw) is certainly possible. A journaled file system shouldn't save 
> you 
> backups

Damn. ;-)

> but the time it needs to recover from system failures. It's not 
> even safe against hardware failures :-)

No! My world breaks apart. ;-)

> > A central place where all the existing jobs are entered is needed 
> > in 
> > any case. Whether it should also be the common place for logging, 
> > well, 
> > I don't know -- I don't think, I understand the whole thing good 
> > enough 
> > to make some qualified statement at this time. :-)
> 
> My plan is like this (for R2, anyway): I will first implement the 
> easiest way to get this thing going, and that'll be the spare 
> partition 
> approach.
> Later, I will investigate in other, more appealing solutions (to Joe 
> user including Tyler and me ;-)).
> I think we have found enough solution and hints for the problems the 
> will come up, that it's time to start developing it. Unfortunately, 
> we'll have to wait a bit until it happens, but that's not sooo bad, 
> since we get a whole working OpenBeOS during that time :-))

Yes, I agree. Let's get something working implemented first, roughly 
keeping possible future extensions in mind.

CU, Ingo