MathWorks at MERGE 2016: Administering Complex Distributed Perforce Environments
Here at MathWorks, introducing edge servers with broker routing capability played a pivotal role in scaling our development. Setting up a local server that handles 90 percent of the load from remote offices has dramatically freed up the commit server and distributed the load among edge servers, considerably improving the overall performance of the remote offices.
With two exceptions, our edge servers are all in one data center. To administer distributed Perforce successfully, we needed to fully understand how routing works.
My talk at MERGE 2016 will address how to administer complex distributed environments and the following issues:
- Decreasing confusion when users’ clients are on different servers
- Using shelved changes in a distributed environment
- Using broker filters to expand the scope of the –df option (p4 client, p4 change, p4 shelve, p4 stream)
- Unloading clients and routing caveats
- Handling controlled failovers in a distributed environment
I hope to see you at the conference!