Client Automation - Improve Stage performance against multiple Scalability Servers
Article ID: 194321
Updated On:
Products
Issue/Introduction
When performing a Stage against a single Scalability Server, it takes few minutes, but when the same package is staged against multiple Scalability Servers on a single job, the time increases exponentially.
Environment
Client Automation - All Versions
Cause
By default each DTS Slave is performing 30 transfers, it sends a parcel to each target and once acknowledged it, they sends the next to all 30 and so on.
The net result of this is, if the Scalability Servers are connected under mixed network speeds, the group of 20 will only transfer as fast as the slowest link.
Resolution
There are different ways of improving performance, but the easiest way is to use the following configuration settings. These settings are only read when the transfers are created (then next time a Stage of a package is made against any Scalability Server).
On the DM through common configuration, Set:
Data Transport Service/Data Transport Agent/MaxConcurrency=50
Data Transport Service/Transfer Object Server/Multi_Group_limit = 5
Data Transport Service/Transfer Object Server/Nos_Addr=NONE
Software Delivery / File Transfer/ DefaultMaxParcelSize=8192000
Feedback
