@orangejulius I've seen a few reports coming through about missinglink/pbf running out of memory when run on large extracts.
unfortunately, this is by design and it's intended that the Valhalla extract method is used for larger extract.
that's not very clear when using the docker setup because it bundles missinglink/pbf in with this code.
merging this will have two benefits:
it will bring in the memory optimizations from https://github.com/missinglink/pbf/pull/15 which AFAIK haven't been released in a latest image for
this repo yet (22 days vs. 18 days)
it will spit out a very ugly warning message if the extract is over 1GB (we can change this constant), so that it will be in the logs just before any out-of-memory error is printed.
@orangejulius I've seen a few reports coming through about
missinglink/pbf
running out of memory when run on large extracts.unfortunately, this is by design and it's intended that the Valhalla extract method is used for larger extract.
that's not very clear when using the docker setup because it bundles
missinglink/pbf
in with this code.merging this will have two benefits:
latest
image for this repo yet (22 days vs. 18 days)