[Gluster-users] Advice in designing a distributed GlusterFS architecture

[Germán Latorre]

  Hello,

We are a company that need a scalable storage architecture for large 
amounts of files, and we are thinking of use a solution based in 
GlusterFS.  We need an architecture consiting in several server files 
configured as a distributed file system that provide HA storage.

I would like to get some advice about an initial hardware and 
architectural configuration that would suit best the GlusterFS needs to 
build a solid system, such as how many servers are needed (we want no 
single point of failure, and have a good I/O performance), and which 
kind of hardware (which kind or servers, disks, network, direct storage 
devices, ...) is needed.  We would start with several TB (maybe around 
20TB?), but our storage system should progressively be able to scale up 
to more than 1PB.

I don't know if this list is the right place to post this message.  
Apologies if it isn't.  Thank you very much in advance for your 
attention in any case.

Best regards,

Germán Latorre.