On 25/02/2016 04:24, Japheth Lim wrote:
On 25/02/16 12:20, Matthew Fernandez wrote:
I think the salient points at this stage are the following:
- Poly/ML performs seeks on the save path to rewind the FD and
update metadata including byte offsets of other sections in the file. Here I'm referring to SaveRequest::Perform.
This is not a problem because segments can be compressed separately. Seeks are only performed to metadata sections and segment starts, as far as I can tell. The vast majority of the data, within the segments, can be compressed without issues as they are not fseek targets.
I had experimented with segment compression earlier, but ran into mysterious segfaults and gave up.
That's how I would have approached it. Is there any chance you could make your experiment available? It may be possible to find out what the segfaults are about. At the very least it would be something to build on.
David