Fedora Servers :: Setting Up A Cluster For Scientific Computing?
Mar 2, 2010
I want to set up a cluster for scientific computing (mainly statistical stuff with R). I have a few conceptual questions. First, is there a difference between a Beowulf cluster and a cluster that has single-system image ("SSI," e.g. using openSSI or LinuxPMI)? If so, what's the difference? Second, if there is a difference between Beowulf and SSI, which one is better for scientific computing? Third, does using Eucalyptus make sense for scientific computing or is this more suitable for IO-oriented operations such as web service or databases
I am bit curious about Cluster Computing. I don't know much about it.Here is Some thing which I wanted to know.1) How can I configure fedora for Cluster computing .2) Is it possible to connect computer of different configuration(like AMD + Intel )? and also for different architecture ( like i386 + x86_64 ) for cluster computing.
I am trying to build GFS2 cluster with 2 or 3 Fedora 14 nodes, but I've encountered some problems from the start. First luci does not work at all in Fedora 14. There is no luci_admin and even if I manage to start luci service, I get a blank white screen when I try to open it from the browser. I've googled a bit and found that I'd might be able to setup GFS if I manage to build cluster.conf manually and start the cluster suite, but I cannot find documentation on how to create cluster.conf anywhere. If anyone knows how to setup GFS2 without a cluster suite or how to configure cluster.conf.
I am working in a project that needs to set up an Apache Web Cluster. The cluster needs to be High-availability (HA) cluster and Load-balancing cluster. Someone mentioned the use of Red Hat Cluster Suite, but, honestly, I can't figure out how it works, I haven't been able to configure it correctly. The project currently have two nodes, but we need to support at least three nodes in the cluster.
I'm trying to build a Fedora 12 cluster to run DHCP on top of two servers using pacemaker and drbd. The pacemaker runs great but i'm having the hardest time getting DRBD installed.
My kernel version is 2.6.32.11-99.fc12.i686.PAE First i tried to [root@dhcp-primary drbd]#yum install drbd this doesn't give me everything because apparently the kernel module for DRBD must be built from source. so i unpacked drbd-8.3.6.tar.gz and tried :
[Code]...
I'm not very experienced in compiling from source i'd love to make an rpm for this but after scouring the web and trying different install methods, I am still at a loss.
Can ANYONE point me in the right direction on how to use storage on multiple servers as a single cluster?I thought storage cluster was for that but, after much googling, and even more help from here, I don't think that achieves my goal. My goal is to have multiple servers share a file system, to act as somewhat of a network raid, so if node-A goes down the files are available on other nodes, and hopefully so when the capacity of the nodes are reached I can add nodes to expand the "cluster".
I need to build a 3 node web server cluster to run a php application. Since the app requires users to login (which means a session state is to be maintained), I will be sharing sessions save path, I also need to share the application directory across 3 nodes. I having trouble deciding which cluster file system to select.
I am trying to install a Red Hat cluster on my servers. In order to configure an heart bit between the servers I must give a valid multicast address. I wanted to know if there is a way to validate a multicast address before using it. For example if it is already in use i don't want to use it.
I read a little about it and there is a Perl module called IO::Socket::Multicast: [URL] There is a tool called mcasttrest.pl: [URL] Oracle use it to test multicast. But I am not sure what will do the trick in Red Hat cluster.
I was following [URL] for the cluster setup. But things dint work out. One Doubt to ask "Do I need the services cman, ccsd and rgmanager one at a time on the both machine. I have been running it through script.
what rpms should I need install for setting up redhat cluster on RHEL 5.0 I want to create two RHEL 5.0 nodes as one cluster having oracle database server installed. And please note I have created these two nodes on VMware server for testing purpose. is it possible for creating cluster of two virtual guests.
I've tried many, many different ways of setting up the cluster.conf file, but every time I start the cman service I get a message telling me "corosync died: could not read cluster configuration".This means nothing to me, nor can I find logs, or anything on the net about this message. I'm ultimately just trying to run a simple GFS2 config on 2 Ubuntu 9.10 desktop nodes, but I can't even get a basic cluster config going.What am I missing? I've been at this for days.
I'm planning on setting up a server cluster specifically for 3D video rendering. In order to maximize speed I wanna use OpenGL hardware acceleration for that and I'm pretty sure that I have to use an NVIDIA video card if I want the whole thing working reliably. Will I be able to start an X-Server with GLX on an NVIDIA video card that doesn't have a monitor connected to it? And what will be the maximum "virtual" display resolution that I can use?
Since I wanna have several servers running side-by-side I really don't have the room for any monitors.Just to avoid misunderstandings: It is not my intention to show what's being rendered to anyone in real time. I will only create video files that can be downloaded later. I'm already pretty sure that this will work, probably using "CustomEDID" or something like that, but I don't have a suitable setup available to test it right now.
how Apache + Ubuntu + Cloud computing works together.Right now, I have [url].... set up to go from my DNS to the cloud server. httpd.conf currently looks like this:
But when you go to 'elliesdev.com' it shows you the EHCP page. I have followed this thread to uninstall EHCP as well as deleting the index.php and its image files from /var/www but I still see the "default web site page" by EHCP. So I'm pretty sure its something with Apache that is causing the site to display that page? What should I do to get ride of it the EHCP default page?I also have a magento ecommerce running at elliesdev.com/magento and I would like to have Apache display it as the default page that you see when you go to elliesdev.com. I have tried changing the DocumentRoot in Apache to /var/www/magento but it makes magento display wrong. How do I make Apache referre to /var/ww/magento as the default for elliesdev.com without displaying wrong?Also, when you go to elliesdev.com/magento, the URL will turn into IP-ADDRESS/magento. How do you make Apache keep my domain name in the URL and not change it to my IP address?I have a feeling that what I am doing is correct, but there is some config file from EHCP that I have not deleted yet that is causing all of this.
I was wondering if and how it would be possible to put together a sort of Home "cluster" system. pecifically, I wanted to make it so I would have a central server and a few computers, where the central server holds the home folders and settings of each user, and each user can log in to any of the computers using a standardized login (into the server), where anyone could log in with their account on any of the few computers - and their home folder (Desktop, Documents, Music, etc) would automatically be there.
Also if possible, would there be some way to centralize the applications that show up on each "client"?
I Installed the 2 disks, but I didn't run into a spot where to place a user name, and when it's time to login, I have a password, but no user name, so I can't login, login fails. And I doubt installing the 2 disks again is going to help this problem go away. How does one login to SL after the install?
Just installed Scientific Linux 6 KDE. Does anyone know how to get it to connect to wired ethernet network?i know there is nothing wrong with hardware; i get a working network/internet connection with live systems such as knoppix, aptosid etc.
I can't forward my kerberos credentials to a computing resource before connecting to the resource for which I have kerberos credentials. In other words, from my machine at work I obtain my ticket with kinit -f to a computing facility off in some lab somewhere.
Then, I want to ssh to another machine in another department (I don't have control over the krb5.conf file or this would have been easy) where I work. It is on this machine I want to be able to ssh,scp, etc to this far off lab. I've tried several options around this barrier, but I'm a total failure thus far. I checked that GSSAPIAuthentication is set to yes.
Install GSL for scientific calculations. When you try to. / Configure appears Permishn denaynd. As I understand it because I have a file system in properties has Read-Only flag. An attempt to change it to read and write an error - saying that this can not be done. All this is done under Root. But nothing came of it, and very much need to put this library.
Fedora 15 uses scidavis (Scientific Data Analysis and Visualization) taken from Fedora 14. If you try to build the native rpm package, an error occurs in building the documentation. A 1-line sed in the spec file solves the problem:
Code: #fix spurious-executable-perm find . -type f -exec chmod 0644 {} ; #fix docbook to adapt different versions of fedora sed -i "s/VER-REL/`rpm -q docbook-dtds|sed "s/^[^0-9]*//;s/.noarch//; s/./\./g"`/" manual/docbook-en/index.docbook # # ---> sed line to fix Fedora15 building: sed -i -e 's/xsl-stylesheets-1.75.2/xsl-stylesheets-1.76.1/' manual/scidavis_html.xsl # #fix default path for fitPlugins sed -i "s/usr/lib/%{name}/plugins\%{_libdir}/%{name}/pluginsg" %{name}/src/ApplicationWindow.cpp sed -i -e 's/Qt;Science;Physics;Math;Graphics;/Education;Science;DataVisualization;Qt/' %{name}/%{name}.desktop
I am new to RHEL. I am using a Thinkpad W510 and SL 5.5. I am trying to get everything working smoothly on my system.I am having issues with the gpu drivers and the touchpad, both of which seem to be traced back to xorg.conf.I can set the proper resolution using nvidia-settings, but it resets to 800x600 on restart no matter what I do. I have used the gui to save, I have tried nvidia-xconfig, and I have tried to just editing the xorg myself, but it always goes back.I am having even worse problems with the trackpad. I downloaded and installed the synaptics package, but nothing I do changes the trackpad settings. Mouse properties has no trackpad info in it.
I'm having a very strange problem with red hat cluster. After testing it in the lab, I tried to install a new cluster and received the following error: cman not started: Cannot start, cluster name is too long or other CCS error /usr/sbin/cman_tool: aisexec daemon didn't start I've checked the internet for that error, but nothing. I decided to to take the example from cluster.conf man, which looks like that :
[Code]...
And still I get the same error. I can find my servers both from DNS (FQDN and short name) and they also appear in /etc/hosts.
i want to install oscar cluster on my ubuntu 8.04 , so i added its repository from http://svn.oscar.openclustergroup.or...ki/repoTesting .i added repo from Testing the development version section.when i use oscar-config --bootstrap i receive this error :
Initializing ODA: mysql ERROR: Unknown ODA type () at /usr/bin/oda line 886[code]....
I am currently trying to setup a cluster using heartbeat. Unfortunatly it does not do what I want to do. I am using heartbeat 2.1.4-11 on a readhat system. I am using the the cib for configuration. What I want to do is the following: I have two Nodes. I want to run a clone resource on both nodes. It's named I want to run as a clone. This works fine.Additionally I want to have two virtual IP adresses. One should be set on Node 1 the other on Node 2. This works fine as well. I want to set dependencies between the resources. And this is where I have problems. What I want to achive: In case named is not able to run on a node, the vIP should switch to the other node (which then will have both vIPs). In case both nodes are able to run named, there should be one vIP on each node. Sounds simple, but does not work.Accourding to the heartbeat docu this an be achieved by using
I trying to create a Cluster Linux with two servers and one storage, I have mounted some filesystems in both servers, but, when I create a file on server (1), I don't see the file on server (2).
Is that a problem with the linux or storage configuration ?.
(Logger pro is an app to retrieve data from scientific calculators/ instruments) I've always wanted to ask this question so here it goes... Has anyone gotten Logger pro (Linux beta) to work w/ Fedora? I have tried alien the two .deb packages to .rpm's so it becomes installable.. The installation seems fine, But the app doesn't run (I think it was like segmentation fault or something... I don't remember) anyways.. this is not absolutely important. I'm just wondering if this may be a problem w/ my install.
Until some time ago I had a dual Windows 7 and Fedora operating system, then I completely moved to Fedora. Recently I have become interested in Scientific Linux and wanted to install it besides my Fedora. I took a back up of all my data and formatted my HDD for a clean install of both of them. This is what I did:
First I installed Scientific Linux with these partitions (contents of /etc/fstab):
[Code]...
Then I went on to install Fedora 15; I kept the same primary and extended partitions but only changed the names of "/FedoraRoot" to "/" and "/" to "/SLRoot" in the above list while installing Fedora. I also chose the boot directory of Fedora to be /dev/sda1/ which I wanted to be the /boot file for both systems while I had chosen the boot file for Scientific Linux to be "/dev/sda".
Everything went well until the final step of copying the files an error came that some of files that Fedora wanted to install in /usr/share, /usr/lib64, /usr/bin, /etc/system..., /etc/issue..., /etc/redhat-release... and /etc/rpm are already present in the Scientific Linux format and so there is a conflict and it has to terminate!!! I must have made some strange mistakes in the definitions
We were also asked to research which is the best and easiest to work with/ setup. So I came here instead of google as no doubt many have used both and have their views of whats what.
I'm tring to setting up some subdomains but it won't work and I don't understand why.I've got the [URL]. Now I want to setup some subdomains for example with test.domain.de I made an entry in my http.conf which looks like this:
