[Tile-serving] [osm2pgsql] remarkable allocation time of flat-node cache (#245)

[apmon]

Yes, indeed, ext3 doesn't support a pre-allocation of file space and the way the allocation system call is handled is for the OS to write a file with zeros to disk of the size you request, which takes quite a while.

With ext4 and some other file system, this call only reserves space in the metadata structures, but doesn't actually write anything in the body of the file to disk, which is thus much faster.

But as Paul sais, for any size of extract where flatnodes is relevant, that time should be fairly insignificant. 

---
Reply to this email directly or view it on GitHub:
https://github.com/openstreetmap/osm2pgsql/issues/245#issuecomment-68384896
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openstreetmap.org/pipermail/tile-serving/attachments/20141230/219702d5/attachment.html>