[PVFS2-developers] PVFS2 Request question
Walter B. Ligon III
walt at clemson.edu
Thu Jul 15 16:41:44 EDT 2004
> Hi Rob,
> So if I use it a sequence of PVFS_Request_contiguous(), it works as
> expected and that is how I am currently using it in the writeback path
> of the ctoo cache( though it is conceivably less efficient).
> Also when the # of tuples is equal to 1 in the hindexed case, it works as
Murali, getting back to our discussion, when you call the PVFS_Request_indexed
and have count set to PVFS_REQ_LIMIT_PINT_REQUEST_NUM the resulting type
will have MORE than that many records, and will be too big.
Try this, run the same program, but set the number of items in your calls
to index to a nice small number, like 10. See what that does. Also,
try replacing the calls with call to PVFS_Request_vector - this will
always be a compact representation, but you can do the same kind of thing
with it. Does that work?
Finally, Rob's question remains - is it a memory type problem or a file type
problem? it would help to know what the erroneous data in the file actually
looks like - is it just shifted over, or totally munged, or what?
Dr. Walter B. Ligon III
More information about the PVFS2-developers