Ask MetaFilter questions tagged with cluster

A beowulf cluster with nothing to do.

brenton — Sat, 12 Sep 2009 15:06:24 -0800

I scrapped all my junk computers together and made the most rag tag beowulf cluster ever in my living room. Mostly as a conversation starter. But now I have to figure out what to have it working on. What's something that will sound really cool?

Does there exist a free cluster/grid, where you get CPU cycles in exchange for providing cycles to the grid?

bsdfish — Sun, 02 Aug 2009 17:59:45 -0800

Does there exist a free cluster/grid, where you get CPU cycles in exchange for providing cycles to the grid? I have come across a reference to Tycoon, which is a software package for operating a grid where people earn credits by donating spare cycles and then can use them to run their own code. That got me thinking about whether there existed a free project that I could join to exchange my free cycles for the usage of the cluster to run my jobs?

There are plenty of the Seti@home type projects that allow you to donate computing credits, but I am clearly looking for computational resources I could use, and I would be happy to pay for with spare cycles. This seems like it would be a very valuable project (many people would be willing to donate a single CPU's worth of credits for a month, in exchange for the use of 30 CPUs in a day), and storage versions of this (Wuala) appear to exist. However, Google doesn't reveal anything useful. Hivemind?

Can I get high availability with a modest budget?

flabdablet — Sat, 20 Dec 2008 21:34:11 -0800

How can I set up the virtual server environment I want, using as much open source and as little proprietary software as possible? I'm the technician / sysadmin at a primary school, which will be updating its on-campus server boxes early next year.

Both our existing boxes run Windows Server 2003. Admin1.admin.local runs SQL Server Express and the school administrative database and serves files to four admin workstations; curricserver.curric.local is pretty much a pure file server for 70 classroom workstations, making extensive use of NTFS permissions to manage access control. All workstations run Windows XP.

Admin1 is the PDC for admin.local (10.145.172.0/23), and curricserver is the PDC for curric.local (10.129.172.0/23). An upstream-managed router connects both subnets to the Internet, and also allows admin.local hosts to send UDP datagrams and establish TCP connections to curric.local, though not the other way around. No broadcast traffic transits the router. The router's link to the admin.local subnet is ten mbits/s, which is fine for internet but sucks for cross-subnet file serving.

Admin1 and curricserver each have a UPS. I use external USB drives for backup. I am pretty happy with the way each of the servers is currently set up.

Next year, we'll also be replacing our existing MS Access-based student reports package with a new web-based one from the same company. I have no reason to believe that this will go well. The company recommends hosting this thing on a dedicated box, which will basically be running SQL Server Express and IIS and not much else. I have no wish to install IIS on either admin1 or curricserver, so I'm happy to agree with them.

Now, rather than buy three new boxes and a new UPS and another set of backup drives next year, I'd rather buy two new boxes with i7's and loads of RAM, run a Linux on them with something like Heartbeat in it, and create the three W2k3 servers I need virtually.

It seems to me that doing this would allow me to (a) keep using the same virtual servers year after year after year, while updating the underlying Heartbeat cluster as often as necessary to suit the school's hardware management policy (b) avoid single points of failure for all servers (c) centralize my backup task (d) use solid Unix system administration tools for disk snapshot and backup management instead of whatever some random commercial vendor claims to have invented this year (e) save the school some money (f) easily bypass the upstream-managed router's connection between our two subnets, allowing me to set up the same safe one-way routing policy at gigabit speeds.

Questions:

1. Is this idea Wrong in any important respect? I haven't had much hands-on VM experience, but I have enough Windows expertise to migrate our existing server setups to other hardware (even virtual hardware), probably without needing to do a Windows reinstall, a long enough beard and enough open source happy drink to see the project through, and the intention to document it thoroughly enough to stop it turning all white-elephant for the next guy.

2. Will Heartbeat in fact let me set up three VM's that will normally run on Tweedledum but reboot themselves automagically on Tweedledee if Tweedledum dies?

3. Which VM environments will let me put a Windows VM in charge of a physical network adapter, so I can serve files from a Windows VM over gigabit Ethernet without undue performance penalty?

4. Am I correct in assuming that I want N+2 physical network adapters in each physical box (one for each of N virtual machines to connect to an appropriate network switch, plus one to do a point-to-point link for DRBD, plus one to talk to the host OS via an appropriate network switch? Or, since the virtual replacement for admin1 is only going to be talking to four workstations, do I actually need a dedicated physical network adapter for that VM?

5. Is there any good technical reason for my visceral unwillingness to install IIS on anything except its own dedicated (physical or virtual) box?

6. What's the obvious question I've completely forgotten to ask?

A Strange Bucket

Blazecock Pileon — Thu, 06 Nov 2008 01:39:58 -0800

I have a bucket containing N marbles: M white marbles and N-M black marbles. I need to grab a handful of marbles (n) and figure out the probability of having picked up m white marbles. At first, I thought I could use the hypergeometric distribution. But there's a complication, namely that the white marbles are not equally distributed in my bucket.

In other words, if my handful of marbles contains one white marble, I'm more likely to have picked up one or more additional white marbles in my hand, and this probability is different depending on how many white marbles I may have picked up.

Is there a good approach to modeling or simulating this kind of situation?

How to cluster when dealing with more than one factor?

Blazecock Pileon — Tue, 30 Sep 2008 02:07:40 -0800

How to cluster when dealing with more than one factor? Let's say I have 100 observations of X, each of length n, with two factors:

1 a_X₁ b_X₁
2 a_X₁ b_X₂
...
n a_X_n b_X_n

1 a_X₁ b_X₁
2 a_X₁ b_X₂
...
n a_X_n b_X_n

(...100 of these guys)

Let's say I have 200 observations of Y, each of length n, with two factors:

1 a_Y₁ b_Y₁
2 a_Y₁ b_Y₂
...
n a_Y_n b_Y_n

1 a_Y₁ b_Y₁
2 a_Y₁ b_Y₂
...
n a_Y_n b_Y_n

(...200 of these guys)

I'd like to calculate the correlation (distance) between X and Y, so that I can cluster them. The two factors may have differing levels of dependence from 1...n.

Is there a general approach for reducing X and Y in such a way that I can cluster them?

Clustering techniques

Blazecock Pileon — Wed, 11 Jun 2008 14:04:02 -0800

What clustering technique to use? I have a set of roughly 500 curves, each curve representing the numerical representation of the behavior of a transcription factor (represented by its binding motif) along a set of genomic coordinates.

In addition, I have six pre-ordained structural classifications. Each of the 500 transcription factor types is a member of one classification.

Presently, I have performed hierarchical clustering of the distances between the curves. I then color the leaves with the according classification, in order to see how the factors organize.

Is there a way to incorporate information from the structural classifications to help assist clustering? What techniques would be better suited for that?

I looked into k-means clustering, but I'm uncertain how I merge the curve information with, say, a six-dimensional unit vector (each axis being the structural classification) that represents membership to a class.

Thanks for your advice.

Large Site Network Infrastructure (Hardware) Examples?

hmaugans — Sat, 16 Feb 2008 16:03:57 -0800

Network admin / geek question: I'm in the process of building a small cluster of servers for my quickly growing wallpaper website, but I've never charted those waters before. I found a nice article on TorrentFreak about Mininova's server setup, and found it quite intriguing. My question is, are there any other big-ish sites like Mininova that are open about their hardware infrastructure? Googling around yielded few results. I'd be interested to see the setup for a site like Digg, MeFi, or Flickr. Thanks in advance. Currently, I have a single Dual P3 load balancer, four DL360 (each 2x 2.8Ghz xeon) servers for apache, one DL560 (quad 2.8 xeon) server for MySQL and NFS, and a few misc. servers.

Do most large sites use NFS to share the website source? Any idea how much load NFS puts on a server? Also, any idea how much overhead a load balancer requires to route connections? Can a dual P3 1.6Ghz handle a 15mbit'ish site?

Any links or advice would be appreciated.

Thanks.

Name our cool computer cluster...

bkudria — Tue, 30 Oct 2007 14:58:19 -0800

Name our cool computer cluster. So a small group of us are building a computer cluster, and we need to name it something ... cool. But, we're completely out of ideas! Please help, hivemind.

Technical details:
14 Dual quad-core Intel Xeon @1.86GHz, with 6Gbps Infinband interconnect, running Debian Linux, with netboot for the nodes.

Guidelines: pretty much anything goes, I think. Puns, parodies also appreciated. Or just something cool-sounding. We probably shouldn't mention Dell or Intel, but other technical references are cool.

LAMP scaling and NFS. Good Idea or Not?

Freen — Fri, 30 Mar 2007 14:41:37 -0800

So, I've got myself a web application. LAMP. I need to scale it up. Can I run the php application across multiple machines loaded off a mounted NFS share on another machine? I'm running gig-e with jumbo frames on a private LAN. Our application is fairly high load, hence the need to scale it up. We can't afford a real fiber SAN, so, can we replicate the benefits of having all of our application scripts live in one place, and each of the web-servers mount the nfs share and serve the application off of the nfs share?

We've thought about other possibilities, such as rsync, etc. But we're concerned about race conditions. Particularly two servers communicating with clients and with the DB using different codebases.

other datapoints:
We use a bytecode cache, called eaccelerator. those caches would be on the local webserver, not on the NFS share.

We use memcache for stuff like sessions across machines.

Would there be significant network throughput setting up our application this way? What are the downsides? Would it outweigh the benefits of having one codebase for our application?

Cool stuff to do with bunch of unused computers?

lifeless — Mon, 19 Mar 2007 02:09:18 -0800

I've got a bunch of HP DC7100 SFF Pentium 4 computers sitting on a table at work, left over from people who have migrated to using laptops as their primary workstations. Any ideas for cool stuff I could do with them? I'm going to a 4-day VMware Infrastructure training later next month, so at that point I'll probably use some of them for experimenting with the capabilities of VMware ESX Server. Another thing I've thought of is experimenting with imaging servers like the ones by Acronis. Both are things that might actually even end up being useful to the employer.

But there's got to be some cool, not necessarily work-related stuff to do out there. I enjoy learning things just for their own sake, so no suggestion, however small the reward may be for the trouble, is automatically out of the question.

I'm a Mac user, though Windows is inevitable at work. I've got an interest in Linux in various server functions (I've got a multi-purpose Linux server in a closet at home), so all suggestions involving Linux (or Linux-to Win/Mac interaction) are welcome. Things like distributed filesystems or clustering come to mind, but I don't really know where to start or where to go with them.

What are some database high-performance clustering options?

SNACKeR — Sat, 02 Dec 2006 05:04:36 -0800

I need recommendations for affordable database solutions that provide high availability and high performance, preferably in a clustered environment. I.e., where you can add another node to the cluster to gain increased performance, without re-architecting your application.

I have been mainly using MS SQL Server for development, which is great for small to mid-size needs, but seems to hit the wall when scaling out. The Microsoft clustering solution provides high availability, but not high performance. I have looked into Oracle RAC, but the cost is hefty (it seems it could easily go over a million dollars for a smallish Oracle cluster). Are there other platforms I should consider that fit in the price gap between Microsoft and Oracle?

Some info: OS platform doesn't matter, the application is a large North American e-commerce OLTP system that will cost the client a lot of money if it goes down. The load is guaranteed to be heavy.

Mysql... why won't you accept more connections?

ph00dz — Thu, 16 Nov 2006 16:20:39 -0800

Mysql5... why won't you accept more connections from a set of IP addresses? Here's kind of a weird one -- we've got a mysql 5 server in a cluster that does a fair amount of i/o. After awhile, it stops accepting connections (or slows way down) from the other machines in the cluster, but if I try to connect from a totally different box, it's lightning fast.

The server itself isn't running a load, nor is it coming close to the number of max_connections/max_user_connection limit. (It is, however, using a single account from all those connections.)

The code connecting to the db is PHP and we do close() the connection.

The mysql itself recognizes that there are "Failed attempts," but it's not clear to us at all what's causing them.

The box is ubuntu, if that helps. The hardware is pretty substantial....

Can the Mac Mini, the VW Beetle of server clusters?

viiviiviivii — Fri, 25 Aug 2006 15:49:48 -0800

I am looking toward upgrading my website's hardware. I want reliability and easy expansion. Could migration to a cluster of mac minis be a good solution for a distributed website? I would be configuring/building the servers in the UK and shipping to their server farm in the US. Basically I was thinking of using mac minis as they are standardised.

My theory is that the cost of adding more servers is relatively low and I will be able to devote new servers to specific sections that need more grunt.

e.g.: Tomcat servers, file servers, database mirrors etc.

Anyway, I have no experience in this field, any advice (with examples) would be appreciated.

DNS behind L4

lundman — Sun, 23 Jul 2006 22:33:04 -0800

Authoritive DNS behind L4? It has been quite a while since I looked at DNS, and I don't recall any particularly difficult parts, but my Manager asked some questions I need to answer with authority.

Assume we host a bunch of authoritive-only DNS servers, say 4. Customer queries, and cache/recursive lookups are done on a different system, and different servers.

It was suggested previously that we could start with a system with an L4 on top, with the external IP. Then have 4 DNS servers on the inside. The content will be identical on all servers in the cluster, running bind. Probably bind with DLZ.

Now there should be no zone transfers required in this setup.

However, if queries are over 512 bytes, the queryers (should) switch to TCP. Would it matter if the UDP and subsequent TCP queries go to different servers in the cluster. Ie, are there any "session keys" or similar content that require the same server both times. (IIRC there is not, and it should just work if the L4 has both UDP and TCP settings).

Will SPF system influence this as well, as it was mentioned.

Nesting vLANs?

Mr. Six — Mon, 26 Jun 2006 13:17:06 -0800

Is it possible to nest virtual LANs within a larger virtual LAN, which has its uplink to the backbone? We have one virtual LAN established in a subnet with its own NAT box and servers inside the vLAN. The virtual LAN has wired outlets in different rooms, and they can all talk to each other and to the "outside" world.

What I would like to do is create a larger vLAN with its uplink to the backbone, and have the smaller vLAN's uplink connected to the larger vLAN.

If it helps to know, the reasoning for this is partially logistic and political:

1. The smaller vLAN would not like to open its services to the larger vLAN.

2. The larger vLAN will be given a direct connection to a Linux computational cluster in a second building. The folks in the smaller vLAN would like access to this cluster.

If I nest the smaller vLAN in the larger vLAN, I figure that I can provide them with access to both the computational cluster and backbone without rewiring the building.

With common switching hardware and software, is this arrangement possible or likely?

Also, have problems with accessing servers within a vLAN from the outside world been solved? Are there ways to configure a server for outside access, or to set up a NAT box to provide outside access to a server?