Synchronize remote studies is slow

[martijnp]

We have 2 PacsOne systems.

PACS1 is running an out-dated PacsOne installation (6.2.3 / Ubuntu 9.04). It contains ca. 9TB of DICOM studies.
PACS2 is a freshly installed PacsOne installation (6.4.2 / Ubuntu 14.4).

We are trying to do a Remote Synchronization of all studies.
- Both sides are connected by a 1Gb/s ethernet card.
- Both AE titles are set to a maximum of 35 connections.
- A flat file 100MB.bin wget over http from PACS1->PACS2 resulted in avg. 820Mb/s

When I check the Live connections on both sides I see that there is only 1 study transferred at a time. This result in a much slower transfer rate.
How can we speed up this process?

I would expect that PacsOne would query multiple studies until the max connections or resource limits are met, so that effectively I would use as much of the 1Gb/s bandwidth as possible.

When I check my resources on the system only 3% CPU and 4% memory is being used.

[pacsone]

PacsOne Server relies on the Dicom Query/Retrive protocol to perform the synchronization with a remote AE, which fetches the Dicom studies in serial (one study at a time) instead of in parallel (multiple studies at the same time).

So if you have an empty database to sync up with a very large remote database (such as the 9 TB database you have) and the remote AE is also running PacsOne Server, it will be much faster to run MySQL Replication to have the empty node synchronization its MySQL database with the existing node, and copy of the raw Dicom image files from the existing node to the VERY SAME storage paths on the new node. The MySQL Replication and raw file system copying are much faster than the serialized Dicom Retrieve, since they utilize the full I/O bandwidth of your server.

[JessicaSn]

I can now confirm, your Apache isn't loading .htaccess files.
You need to set AllowOverride to All in your apache config file and then restart your apache server.

Or copy the rules from the .htaccess in your apache configs and then restart it.