Tag Info

Hot answers tagged

13

Please don't take offense to this but I strongly suggest you bring in a local area IT consulting firm that specializes in systems and network administration. I also came from a programming background many moons ago and learned many hard lessons on the do's and don'ts of managing a networked server environment. I (thankfully) had alot of mentors and help ...


12

There are many methods for scaling web applications and the supporting infrastructure. Cal Henderson wrote a good book on the subject called "Building Scalable Web Sites". It was based on his experiences with Flickr. Unless you grow slowly you will run into the same kinds of growth issues many others have seen. Scaling is, like many other subjects, a ...


12

"uname -m" or "arch"


12

First off, a 2.6GHz processor is not a 2.6GHz processor if they're from different generations. You're correct in thinking twice about that. This has been true for a long time now (at least since the 486 / Pentium days), and so it's important to point out to your engineer just how wrong the Megahertz Myth is. Especially given the massive performance ...


11

The lm flag in your /proc/cpuinfo indicates that your server has a 64 bit CPU. The info from file prog indicates that the program is compiled for 32 bit architectures. Try installing the 32 libraries. I'm not familiar with how to do this in Gentoo, but maybe this Gentoo wiki article can help. I have had a simliar problem (bash reporting the file not being ...


11

You either use Q-in-Q, or yes, you build out a second isolated switch forest with it's own set of VLANs.


9

Best known example is EVE Online, which uses Stackless Python. They have written they own server. Their architecture described: http://highscalability.com/eve-online-architecture http://www.massively.com/2008/09/28/eve-evolved-eve-onlines-server-model/ http://www.slideshare.net/Arbow/stackless-python-in-eve Generally I recommend reading ...


8

If you want to see the "bitnes" of the OS installed you can run this command. $ uname -m Regarding the actual capabilities of the processor you can always look the model up inside /proc. $ cat /proc/cpuinfo


7

It's really just a process of listing what you have, then listing what you need. Then using the first list (your assets) to build out a network that does what you need. You may have to buy things you don't have, to create the network which meets your needs. Like so: First you make a list of the sites. For each site, you make a list of what's there. If ...


7

So this is sort of an answer, but probably not what you want to hear. Don't do it. It may seem appealing at first to have a server "golden image" that "has it all". You don't want to go down that road. Here are a few reasons why: Each library you install provides another possible avenue for security breaches. Each time you clone this server, you'll need ...


7

cat /proc/cpuinfo look for the 'lm' flag, it means 'long-mode' i.e. 64-bit capable.


7

Yeah, it's valid. My personal approach would be to use Varnish up front and use VCL to split the traffic between static NGINX requests and your heavy lifting (whether that be Apache or Passenger or... it doesn't matter). This is especially true if it is on the same machine as you don't need the extra overhead. It doesn't necessarily buy you anything.


7

The controller as a performance bottleneck is quite true, and it can represent a single-point-of-failure as well in some architectures. This has been known for quite some time. For a while there were vendor-specific techniques for working around this, but since then the industry as a whole has converged upon something called MPIO, or Multi-Path I/O. With ...


6

The command on Linux/UNIX is: uname -a or for just the architecture: uname -m


6

From commandlinefu.com: getconf LONG_BIT


6

My favorite one is this: http://blogs.technet.com/wbaer/archive/2007/07/10/understanding-hierarchy-and-basic-concepts-of-navigation-in-microsoft-office-sharepoint-server-2007-windows-sharepoint-services.aspx But some people find that a little too technical. I have also used the following in the past: ...


6

While I generally prefer Linux on the server side, this is definitely not one of those situations. Stick with Windows, and you will never know the horrors of managing Windows permissions with POSIX ACLs. Don't dump files into databases, and your backups will be much easier. Build a clean Active Directory environment, and you will have time to take long ...


6

The term architecture covers a lot more than just the processor. There is a lot of other hardware components that are crucial to an OS kernel. The first example that comes to my mind is the interrupt controller, which is separate from the processor, but depending on actual model may be put inside the same chip as the processor. This entire collection of ...


5

Although network bandwidth comes into play in a big way, the absolute number one factor to consider is what is the transaction log generation rate on the principal? If the app and your maintenance doesn't generate any transaction log, then network bandwidth is really irrelevant. If it does generate log, then network bandwidth has to be able to handle the ...


5

Microsoft published a really good whitepaper on database mirroring that includes some good examples on how much performance impact you get from synchronous mirroring. You're totally right in that there's going to be a performance hit. Do a ping from the primary box to the database mirror and look at the round trip times in milliseconds: that's going to be ...


5

It makes no difference, inbound routes won't be considered when assigning reputation. Reputation used by email spam filtering providers (one of whom I work for) is based entirely on the messages that that IP address sends, not what it receives (we can't see that anyway). Nor would using different IP addresses cause a problem for an MX filtering provider, ...


5

It's virtual address space per process, as per the MSDN article, and the superb series of articles on this written by Raymond Chen and archived at his blog. Here is his index page for this series of articles - very well worth a read if you're dealing with large memory support as a senior system admin or a developer.


5

You can use RRAS for firewalling, NAT and VPN, so, yes, you can give a single public IP address to your Windows Server 2008 firewall and have it route traffic for all your internal network and forward specific ports (f.e. 80) to your internal servers, and you can also have it act like a VPN server (PPTP and/or L2TP). RRAS has been around since Windows 2000, ...


5

(I was just about to submit the answer on stakoverflow when you moved it to ServerFault :-) ). The basic idea is that you want to have different servers optimized to do one thing very well. I.e. if you have an application server running code / storing state, etc, you want to optimize your hardware for that purpose - i.e. good CPU and lots of memory. On ...


5

Varnish doesn't (yet) support gzip compression, so it might be an idea to swap it around with nginx in front to compress what varnish sends back. Since varnish and nginx don't fight for the same resources (nginx uses CPU for gzip compression, while varnish uses memory) they should run smoothly on the same machine. Varnish now supports gzip compression, so ...


5

One useful benefit of "L7" like haproxy is being able to use cookies to keep the same browser hitting the same backend server. This make debugging client hits much easier. L4 balancing may bounce a single user around on several backend servers. (which in certain cases may be advantageous, but in a debugging/profiling sense, using "L7" is much more ...


5

Try out http://highscalability.com/display/Search?searchQuery=google it's full of resources.


5

If you really want your application to run as it should against the different versions you want them cut off from each other - so I'd advise you have 3 x XP installations. If you use ESXi 4.1 you'll see some significant memory savings between the 3 VMs so I wouldn't worry too much about that. Essentially if you're going to do something do it well :)


5

This is a pretty loaded question. My general advice is to focus your attention on managing complexity and allow the system to grow organically. Virtualization: You really want to avoid server sprawl, and these days, everything is virtualized. Pick a platform that will allow you to add virtual servers quickly, as well as manage them efficiently. One trend ...


4

I second Evan Anderson's suggestion of MPLS. It's the solution that I have considered for my own network. Logically, it's point to cloud, and all the complex routing is done by the provider. This necessitates that you have one unified provider across your entire infrastructure*. This isn't as bad as it might sound, since there's a bit of configuration ...



Only top voted, non community-wiki answers of a minimum length are eligible