From: Anthony G. <an...@fa...> - 2008-11-22 02:06:52
|
Jonathan, I have not yet tested the maximum throughput of a single php Thrift client. We use Scribe to collect data from thousands of machines and are primarily concerned with scaling the aggregate total amount of data. We are usually limited by how fast we can write to disk, but can easily scale out by configuring scribe to forward messages to as many scribe servers as we need. The exact throughput would depend on the workload and how Scribe is configured. I am not sure what you mean by 'configuring php on scribe', but I am assuming you are asking how to configure Scribe. If so, I would take a look at the configuration section of the Scribe wiki. Also, I recommend taking a look at the example code that is included with Scribe. The example code demonstrates how configure Scribe to run on multiple machines and buffer data locally when remote Scribe instances are not available. -Anthony On 11/20/08 5:49 PM, "Jonathan Cao" <jon...@ro...> wrote: We are trying to evaluate Scribe log server. What is the typical throughput for php Thrift client? Any advice to configure the php on scribe would be welcomed. Thanks, Jonathan |