Latest version performance issues

Urbackup Server: 2.5.31 on Ubuntu 22.04.3
Server Settings: /etc/default/urbackupsrv

Defaults for urbackupsrv initscript

sourced by /etc/init.d/urbackupsrv and /lib/systemd/system/urbackup-server.service

installed at /etc/default/urbackupsrv by the maintainer scripts

This is parsed as a key=value file

#Port for FastCGI requests
FASTCGI_PORT=55413

#Enable internal HTTP server

Required for serving web interface without FastCGI

and for websocket connections from client

HTTP_SERVER=“true”
INTERNET_ONLY=true

#Port for the web interface
#(if internal HTTP server is enabled)
HTTP_PORT=55414

#Bind HTTP server to localhost only
HTTP_LOCALHOST_ONLY=false

#Bind Internet port to localhost only
INTERNET_LOCALHOST_ONLY=false

#log file name
LOGFILE=“/var/log/urbackup.log”

#Either debug,warn,info or error
LOGLEVEL=“debug”

#Temporary file directory

– this may get very large depending on the advanced settings

DAEMON_TMPDIR=“/tmp”

#Tmp file directory for sqlite temporary tables.
#You might want to put the databases on another filesystem than the other temporary files.
#Default is the same as DAEMON_TMPDIR
SQLITE_TMPDIR=“”

#Interfaces from which to send broadcasts. (Default: all).
#Comma separated – e.g. “eth0,eth1”
BROADCAST_INTERFACES=“”

#User the urbackupsrv process runs as
USER=“urbackup”

NGINX conf: urbackup.conf

# Make UrBackup webinterface accessible via SSL

server {
# Define your listen https port
listen 443 ssl;

    # (optionally)
    # server_name urbackup.yourdomain;

    # SSL configuration
    #ssl on;
    include ssl-params.conf;
    ssl_certificate /root/.acme.sh/server_ecc/fullchain.cer;
    ssl_certificate_key /root/.acme.sh/server_ecc/server.com.key;
    # SSL configuration

    # set the root directory and index files
    root /usr/share/urbackup/www;
    index index.htm;

    # This location we have to
    # Proxy the “x” file to the running UrBackup FastCGI server
    location /x {
       include fastcgi_params;
       fastcgi_pass 127.0.0.1:55413;
    }

    # If come here using HTTP, redirect them to HTTPS
    error_page 497 https://$host:$server_port$request_uri;

    # Disable logs
    access_log /root/access.log;
    error_log /root/error.log;

}

I had speeds of 150-800Mbit/s for backups from Windows and Linux clients over the Internet depending on the remote client Internet connection.

After the upgrade I am seeing 30Mbit/s. I have tried changing all kinds of settings, but nothing seems to improve the speeds.

The Server is quad CPU 80 cores total, 265GB RAM and BTRFS RAID10 250TB of SAS storage. I was going to put two 4TB of caching in-front of it, but my testing shows speeds seemed to be fine without it.

I have about 5TB backed in BTRFS now. The system stats with backups running:

IO is: .3
RAM is: 4G used. so tons of ram available.
CPU is: 99.9 free so .1 is being used
5G Fiber Internet direct to the box. 10G Total Fiber. iperf3 shows network is performing as it should to remote and local systems.

SQLite database checks out fine (is there a document for moving to MySQL). I tried to disable Encryption and Compression no change in speed. Increased the file threads, hash threads to 40. Just not sure why I am not getting the performance I saw before.

Thanks in advance!