| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Having a shared temporary directory, flock(2)'ed to avoid races, is a
great idea in theory but unfortunately doesn't work so well with
systemd.exec(5)'s ReadWritePaths settings since
ReadWritePaths=/var/www/webmap/tiles
ReadWritePaths=/var/www/webmap/tiles.tmp
creates multiple mount points pointing at the same file system and
rename(2)/renameat2(2) can't cope with that. Quoting the manual:
EXDEV oldpath and newpath are not on the same mounted filesystem.
(Linux permits a filesystem to be mounted at multiple points,
but rename() does not work across different mount points, even
if the same filesystem is mounted on both.)
So the options are to either use a single ReadWritePaths=/var/www/webmap,
or --mvtdir-tmp=/var/www/webmap/tiles/.tmp. Both kind of defeat the
point (we'd in fact want to use --mvtdir-tmp=/var/tmp/webmap/tiles), so
we use mkdtemp(3) instead.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
It's done automatically by the MVT driver. Also we don't want to clip
at the extent boundaries, but instead leave som headroom so the clipped
polygon border is not visible on the map. The MVT driver does that.
It seems that GDAL 3.6.2 from Debian Bookworm generates too many tiles
though. It yields the following tile counts for group ‘ren’:
no manual clipping, BUFFER=32: 83718 tiles [min=33 B, max=117.70 kiB, sum=15.73 MiB, avg=197 B]
no manual clipping, BUFFER=0: 83676 tiles
clip at extent, BUFFER=32: 76256 tiles
GDAL 3.10.3 from Debian Trixie yields less surprising tile counts:
no manual clipping, BUFFER=32: 75972 tiles [min=33 B, max=128.16 kiB, sum=15.10 MiB, avg=208 B]
no manual clipping, BUFFER=0: 75939 tiles
clip at extent, BUFFER=32: 75972 tiles
(Interesting to see that the largest tile — 0/0/0.pbf — is over 10kiB
larger with the more recent GDAL version, also.)
|
