[Pvfs2-developers] server crash on startup with millions of files

[Robert Latham]

On Tue, Feb 20, 2007 at 12:23:56PM -0500, Phil Carns wrote:
> The 5 minute scan time is a problem, because it makes it hard to tell 
> when you will actually be able to mount the file system after the 
> daemons appear to have started.  We would be happy to try out any 
> optimizations here :)

The original design of the handle managment was supposed to keep the
stateof used/free handles on disk, but I couldn't come up with a good
way to do this efficently.  Neill and I did this scanning thing, "just
to get things working", which of course ended up sticking around for a
while :>

Perhaps it's time to revisit the handle management code (I believe
it's the oldest PVFS code left in the tree).

If a quick read-and-discard of the trove db files in trove
initialization means we can deal with millions of files in a few
seconds, maybe that's a sufficent (if kind of gross) hack. 

==rob

-- 
Rob Latham
Mathematics and Computer Science Division    A215 0178 EA2D B059 8CDF
Argonne National Lab, IL USA                 B29D F333 664A 4280 315B