"Aaron J. Grier" writes:
> On Thu, Jan 18, 2001 at 08:05:34AM -0600, Greg Oster wrote:
> 
> > So even allowing fsck and the parity rewrite to happen at the same
> > time is stretching things a little...
> 
> ouch.  so there really should be two sections of /etc/rc.d/raidframe:
> a pre-fsck configure, and a post-fsck parity check and rewrite.
> 
> when can I expect a pull-up for 1.5.1?

Ummm.... the parity check should *really* be complete before the 
fsck even begins.  I'm somewhat uncomfortable even having the parity re-write 
being done in the background, as there *is* the chance of data loss if one
of the drives fail before the re-write completes.  But people seem to want
their machines up before the 2 hour parity checks complete, so doing it 
the background seems to be 'acceptable' for most people.  I'm not at all 
sure, however, that we want to delay the parity check until after the fsck
completes... (IMHO the solution here is to reduce the amount of parity 
checking required, not to increase the window in which a component failure 
could be catastrophic...)

Later...

Greg Oster