Home › Forums › General Discussion

Optimizing heavy read environment

Mon, 09/03/2007 - 23:55 — Anonymous

Hi,

Our situation is a simple one:

2 storage servers, each with 2 bandwidth feeds in and then a BGP daemon on each server handling routing (so we have some control over cost).

Each server has 2 terabytes of storage, at any time 70% between the 2 servers is in use. There's probably 2mbps of write going on at any one time and the read is the issue. There is no issue with bandwidth, each server has 2gbps to it if it needs it (we run the network so we know there's no bottlenecks). The problem is, even after stopping all http services and then booting them back up we get pretty dire performance. We use Nginx to handle the downloads. We're sure it's not hardware but something that's fixable in software and we're just overlooking.

Setup is

3.2ghz xeons with 3ware raid cards, each server having 4 disks running freebsd + nginx. SQL is on an external server, each server has 4Gb ram.

Any help is appreciated

Thanks

‹ Technology behind mediatemple grid service 2 tier switch selection for colocation ›
  • 1250 reads

Post new comment

The content of this field is kept private and will not be shown publicly.
  • Web page addresses and e-mail addresses turn into links automatically.
  • Allowed HTML tags: <a> <em> <strong> <cite> <code> <ul> <ol> <li> <dl> <dt> <dd><div ?=?><p ?=?> <img ?=?><h1 ?=?><h2 ?=?><h3 ?=?>
  • Lines and paragraphs break automatically.
  • Glossary terms will be automatically marked with links to their descriptions
  • You may link to webpages through the weblinks registry

More information about formatting options

To combat spam, please enter the code in the image.