PUBLIC MARKS with tag cluster

"Coda is an advanced networked filesystem"

Naming: Marvin is named after Marvin, the paranoid android from the Hitchhiker’s Guide to the Galaxy. Marvin has “a brain the size of a planet”, which he is seldomly able to use. Indeed, the true horror of Marvin's existence is that no task he could be given would occupy even the tiniest fraction of his vast intellect.

CMON is a daemon that aggregates information from MySQL Cluster that earlier was only accessible from the cluster log or the management client, such as: * cluster state * node state * backup statistics * statistics * cluster events (cluster log basically) .. and let's you access the information using SQL, because CMON logs the information into ordinary MYISAM tables! So, it is really easy to use! In the package you also get get php scripts that you can put on your webserver to generate graphs and get a www interface to CMON. CMON can also start ndbd nodes and make decisions on how they should be started (with or without --initial). CMON starts as a daemon and will automatically create cmon database and install the necessary tables automatically.

In order to be able to do a live migration of a Xen guest from one cluster member to another, some sort of shared storage is required. As the Xen guest won’t run on more than one cluster member at a time, a cluster filesystem is not required. That is, as long as you configure Xen to access the Xen guest by a physical device, not a file.

Quand un site web dynamique commence à avoir un trafic important, généralement on va essayer de multiplier les serveurs web qui hébergent les fichiers. Multiplier les serveurs web n’est pas le plus difficile, il suffit de faire une réplication des données à chaque mise à jour de votre site et de multiplier les sous-domaines. Les choses se compliquent lorsque vous souhaitez avoir plusieurs serveurs de base de données. La base de données que nous étudierons ici est MySQL. Nous ne parlerons pas de cluster MySQL dans cet article. Le but de l’article est de montrer comment séparer un serveur MySQL en trois serveurs distincts : 1 maitre et deux esclaves, avec un seul point d’entrée (le proxy MySQL). Le proxy devra envoyer les requêtes de lecture vers les serveurs esclaves et toutes les autres requêtes vers le serveur d’écriture.

Many people use SSH to log in to remote machines, copy files around, and perform general system administration. If you want to increase your productivity with SSH, you can try a tool that lets you run commands on more than one remote machine at the same time. Parallel ssh, Cluster SSH, and ClusterIt let you specify commands in a single terminal window and send them to a collection of remote machines where they can be executed. Why you would need a utility like this when, using openSSH, you can create a file containing your commands and use a bash for loop to run it on a list of remote hosts, one at a time? One advantage of a parallel SSH utility is that commands can be run on several hosts at the same time. For a short-running task this might not matter much, but if a task needs an hour to complete and you need to run it on 20 hosts, parallel execution beats serial by a mile. Also, if you want to interactively edit the same file on multiple machines, it might be quicker to use a parallel SSH utility and edit the file on all nodes with vi rather than concoct a script to do the same edit. Many of these parallel SSH tools include support for copying to many hosts at once (a parallel version of scp) or using rsync on a collection of hosts at once. Because the parallel SSH implementations know about all the hosts in a group, some of them also offer the ability to execute a command "on one host" and will work out which host to pick using load balancing. Finally, some parallel SSH projects let you use barriers so that you can execute a collection of commands and explicitly have each node in the group wait until all the nodes have completed a stage before moving on to the next stage of processing.

I have tried to run both Debian Etch and Ubuntu 8.04 Server on the cluster nodes, in Dom0. I started my tests with Debian, but I had some issues with slow samba performance in one VM that I couldn't fix so I decided to try Ubuntu Server, for the first time. Both installation went OK, the main difference was that I used mainly source code in Debian, but only packages in Ubuntu. I actually ran into more problems with Ubuntu due to some early bugs in the 8.04 release, will describe them below as I go along. And I have still to prove that running this setup in Ubuntu is stable.

smtp-delay is an add-on/plug-in intended for use with qmail. It was written primarily to add banner delays and antipipelining to qmail. These two features are known to be able to block certain types of spam and virus mail sent through non-rfc-compliant SMTP engines. When I looked around for programs to add this functionality to qmail, I found only one such program, and didn't like the way it was done. BTW...I have the same objections to the way its done in sendmail 8.13.x. Since banner delays (the server pausing for some time before issuing an SMTP banner) cause every SMTP connection to take longer, I thought it would be a good idea to somehow exempt "legitimate" mail servers...or at least not subject them to long banner delays. So I decided to tune the banner delay time based on the connecting IP's reverse DNS. IPs with no rDNS get treated the worst (longest banner delay). IPs with rDNS matching a regex intended to detect dynamic/end-user IPs get a moderate delay. All other IPs get a very short banner delay...just long enough to see if they immediately pipeline (send SMTP commands before the banner's been sent). The original intent for smtp-delay was that it should be run before rblsmtpd, and simply set the RBLSMTPD environment variable if applicable, letting rblsmtpd issue the 4xx response. Pretty early on, I realized smtp-delay should be able to run standalone (without dependence on rblsmtpd to do its talking) and issue a 4xx response on its own. Lately, the spam load against our mail cluster has gotten so bad that I've started running smtp-delay after rblsmtpd, based on the idea that there's no point waiting out a long banner delay holding an open socket to an IP we have no intention of accepting mail from anyway. This reduced our concurrency by about 20%.

OpenNebula transforms a physical cluster into a flexible virtual infrastructure which dynamically adapts to the changing demands of a service workload. OpenNebula leverages existing virtualization platforms to create a new virtualization layer between the service and the physical infrastructure. This new layer supports the execution of the services on a physical cluster, extending the benefits of VMMs (Virtual Machine Monitors) from a single physical resource to a cluster of resources. OpenNebula effectively decouples a server (deployed as a pre-configured VM) not only from the physical infrastructure but also from its physical location.

Bigtable-like structured storage for Hadoop HDFS

By default, migrating a Xen Virtual Machine (VM) resource causes it to shutdown on the current node and restart on the new one. Once you configure a Xen VM within the High Availability Storage Infrastructure (HASI), how do you configure the Xen VM resource to live migrate among the cluster nodes?

This tutorial is based on Falko Timme's tutorial for MySQL Cluster 5.0. It shows how to configure a MySQL 5.1 cluster with five nodes: 1 x management, 2 x storage nodes and 2 x balancer nodes. This cluster is load-balanced by an Ultra Monkey package which provides heartbeat (for checking if the other node is still alive) and ldirectord (to split up the requests to the nodes of the MySQL cluster). In this document I use Debian Etch 4.0 for all nodes. Therefore the setup might differ a bit for other distributions. The two data nodes were x64 to use all of the 8GB RAM. Servers were compiled from source so you should be able to make it running on any platform. The MySQL version I use in this setup is 5.1.24-rc. It's a release candidate, but I wanted to use 5.1 to take advantage of Memory-Disk Based tables. Beginning with MySQL 5.1.6, it is possible to store the non-indexed columns of NDB tables on disk, rather than in RAM as with previous versions of MySQL Cluster.

This tutorial shows how to set up a two-node Apache web server cluster that provides high-availability. In front of the Apache cluster we create a load balancer that splits up incoming requests between the two Apache nodes. Because we do not want the load balancer to become another "Single Point Of Failure", we must provide high-availability for the load balancer, too. Therefore our load balancer will in fact consist out of two load balancer nodes that monitor each other using heartbeat, and if one load balancer fails, the other takes over silently.

The sfSyncClusterPlugin introduces a symfony sync-cluster task. The symfony sync-cluster task is intended as a replacement for the symfony sync task. While the plugin is backwards-compatible with symfony sync (they even use the same configuration files) this plugin offers several advantages over a traditional symfony sync. The symfony sync-cluster task: * Synchronizes a Symfony application across an unlimited number of servers. Symfony's sync only works for a single server. * Clears the Symfony cache on each server with a symfony cc. Doing this with sync requires an additional command. * Allows you to use ssh public/private keyfile authentication so you won't be prompted for a password during long deployments. * Allows you to deploy your web assets to a CDN without having to copy your entire site. The sfSyncCluster currently supports standard rsync transfer and transfer to Amazon S3. * Allows you to specify groups of servers to sync. If you have many servers, you can place them into logical groups and move them around quickly.