The only reason to increase the scaling up performance is when you have 32-48 CPU cores and you wish for your file to be processed very fast. In order to increase the scaling up performance, there are 2 approaches as follows:
If you have a Scan Node (formerly known as Agent) connected to Core and the queue is full of incoming scan requests, they will be denied until there is a free slot.
To change the queue size, you can use the following API call:
PUT on http://<serveraddress>:8008/admin/config/scan
With the following body:
Note that to use this API you need an apikey,
apikey can be obtained from the Core Management Console→Settings→User Management→Admin user
use the apikey in the apikey header.
This works for both Linux and Windows.
Another way to better utilize your CPU, if you have 16-32+ cores, is to increase the parallel number of scans running per engine. This way you can tell the Node how many scans should run simultaneously per engine. This applies to engines supporting multi-threaded scan.
To change this you have to write in the node's config in the registry:
HKLM/Software/OPSWAT/Metascan Node/internal/parallelcount (or HKLM/Software/OPSWAT/Metascan Agent/internal/parallelcount in case of versions <= 4.5.1)
If the internal key is not present just create it and then create the new config value. The default value is 20.
After you modified the value, please note that you have to restart the MetaDefender Core services for the changes to take effect. To restart the services, please open an elevated command prompt and type the following commands:
net stop ometascan
net stop ometascan-node
net start ometascan
net start ometascan-node
(use service name ometascan-agent instead of ometascan-node for versions <= 4.5.1)
On Linux, you have to modify Node's startup configuration file /etc/ometascan-node/ometascan-node.conf by adding "[internal] parallelcount=the new config value" and restart the service.
This article applies to MetaDefender Core v4 Windows and Linux
This article was last updated on 2019-10-06