unexpectedly low performance

Hello, i’ve trying to diagnose low store performance on orthan for more then a week ago,

Currently i’m managing a server that is storing about 2.5M instances and 1tb of data in total, it was running sql lite database at first and was seeing writes about 1MB/s, then switched to postgresql and they went up to 1.5MB/s typical.

its running on a freebsd jail on a truenas system. first i cheked if the drives where the fault and installed it on another pc running baremetal freebsd with default config and sql lite and one of the drives,
got around 1-5 MB/s stores, negligible cpu and ram ussage, then proceded to run it on ubuntu and debian, same speeds.

if in any of those systems do a regular smb file transfer it goes around 50MB/s.

then proceded to run it on an ssd with the default config and sql lite, only then did i see a bump in speed but it was to 8MB/s. and still negligible cpu and ram usage.

what could cause this? what other poeple get on transfer speeds? because on the clinic the server is operating the workstations and a disk publisher can push up to 500Mb/s network transfers on dicom files.

FYI, here are some performance tests I ran 2 years ago to compare object-storage performance with SSD: https://bitbucket.org/osimis/orthanc-setup-samples/src/master/docker/performance-tests/

Tests were done through HTTP but DICOM protocol should be close in terms of performance.

HTH

Alain

Thanks for the info
Also and update, ran orthanc using docker on a manjaro system, and got the speeds I would expect, around 20-50 MB/s still don’t know how the freebsd port is so slow or even the packages from ubuntu