Dear List, I've been way behind on versions (happily using fuseki 2 for a long time). But with log4j, I'm trying to switch over to the most current version of Fuseki (4.3.1).
Initially, I wanted to just use the TDB build that I had ingested using fuseki 2 and load that using fuseki 4. But that didn't seem to work. So I decided to re-ingest everything into a new TDB build using fuseki 4.3.1. This seemed to work better. However I am noticing a dramatic difference in space usage. The size of my old TDB build (created using fuseki 2) was 51 Gigs. The size of the new TDB build (created using fuseki 4) was *152 Gigs.* There are about 30,892,337 in each build Does this large jump in disk size make sense as part of the move from fusek 2 to fuseki 4.3.1? For example: Is 4.3.1. creating new indices or other kinds of ancillary data that could account for the large jump in disk space used? On the surface it doesn't make sense to me and it suggests I have a problem elsewhere, but before I start over and look for other problems I wanted to make sure this jump in disk space usage wasn't something I should expect from moving to fuseki 4.3.1 from fuseki 2. Here are some break-downs of individual file sizes in the two builds In the previous (50g build from fuseki 2) 792M Aug 3 18:51 node2id.dat 112M Aug 3 17:20 node2id.idn 724M Aug 3 18:51 nodes.dat 0B Aug 3 18:51 nodes.dat-jrnl In the new (150g build from fusek 4) 16B Dec 14 06:07 nodes-data.bdf 650M Dec 14 04:19 nodes-data.obj 24B Dec 14 06:07 nodes.bpt 69G Dec 14 04:19 nodes.dat 10G Dec 14 04:19 nodes.idn The additional 79G in nodes.dat and nodes.idn could almost entirely explain my near 100G increase Another large increase is in the POSG file: In the 50gig build this was: -rw-r--r-- 1 jcwitt staff 6.5G Aug 3 18:52 POSG.dat -rw-r--r-- 1 jcwitt staff 456M Aug 3 18:51 POSG.idn and in the new build this was: -rw-r--r-- 1 jcwitt staff 24B Dec 14 06:07 POSG.bpt -rw-r--r-- 1 jcwitt staff 23G Dec 14 04:19 POSG.dat -rw-r--r-- 1 jcwitt staff 13G Dec 14 04:19 POSG.idn and near 33gig increase. Combined with the above 79gig increase in the nodes file, this would seem to explain it. Any ideas why this is happening? Many thanks for any generous assistance you can provide.
