Geographically Distributed MySQL

Greetings to all. There is a desire to geographically distribute the project, and start with one of its components: MySQL. Interesting answers are those who worked closely with this database and not in theory knows how various geographically distributed balancing schemes work. The current scheme is approximately as follows: one web server and two database servers in master-slave mode. Only read requests go to one, mainly to write to the other, both database servers are nearby and are con... Read more

Clustering - what should be

This is again with link link link. What should include cluster management software? I’m not in the subject myself, so I’m hoping for cloud gurus. What is now: * Switching the entire panel to the cluster configuration mode - editing a separate configs tree is activated. * All cluster configs are stored as an overlay for root filesystem * Ability to substitute variables in any place of any config and set them separately for each node: * It stores information abou... Read more

Choosing a cluster file system?

Advise a cluster FS (that is, a FS that can be mounted on different nodes in rw), which could combine the stack with a hundred terabytes. You also need FS to use the native kernel of CentOS / Debian. For this reason, GPFS and Luster are unfortunately not suitable. I look in the direction of the GFS, but I strongly doubt that it can evenly distribute the load evenly across the physical media. As I read, filling the data goes there linearly. Read more

Do I need to organize a thematic meeting? Does anyone go?

There was an idea to organize a meeting of system administrators on the topic “building a cluster for shared-hosting”. Of course, I have a personal interest in this topic. This is a task that cannot be solved in a narrow circle of people (in one company, for example). The ultimate goal of a number of activities that will begin with this meeting is to write complete documentation (wiki) about creating an easily manageable cluster, with a convenient panel that is friendly for end users. And, of ... Read more

Database clustering

I read in the network that VKontakte uses MySQL as the main repository. Yes, and many other projects use MySQL. An interesting question is, how is this implemented from the technical side? Not exactly like theirs, but how it can be designed in general, the main requirements: Using free databases Transparent work with the database (i.e., scripts do not need to know how and what is arranged there and connect either to a single server always or to a random one in a cluster) If a single serv... Read more

How and how to merge the server into a cluster?

There are three servers - 8 cores, 16GB of memory, 2TB disks on each. Would you like to integrate them somehow so that they look like one real powerful server? Preferably with Linux. Read more

Do not tell me where to find the lit?

Putyatin EP Image processing in robotics. / E.P. Putyatin, S.I. Averin - M .: Mashinostroenie, 1990. - 320 p. Methods of correlation detection of objects / [A.V. Girenko, V.V. Lyashenko, V.P. Mashtalir, E.P. Putiatin] - Kharkiv: Business Inf., 1996. - 112 p. Or information about classifying the classification in images? Read more

Server clustering of geotag cards?

There is a map of Moscow, there are more than 5,000 tags on it. Of course, they do not need to be displayed as one continuous messiva, but it is necessary, at a certain scale, to merge adjacent labels into one with a number. How to do it through the client side - I know there are libraries, and in Google Maps it even comes in the standard part. And how to do it on the server side? There is a database with 5,000 labels, each of which has a longitude and latitude. How to merge neighboring point... Read more

Does anyone use data clustering services on maps?

In one of the projects, due to the rather large number of markers on the map (~ 300k), it became necessary to cluster them to improve the performance on the client side. From memory, he recalled only two services of claterization: Maptimize With good performance and advanced features to work with the properties of the markers, but with biting prices. Geocubes Nina, apparently already closed. Does anyone use similar services in this area, perhaps with deployment on their servers? It... Read more

Tell me on the topic of Linux HA + shared storage?

Good afternoon, I had a task to build a failover cluster with virtual machines running in failover mode. Cluster as storage should use shared file storage. By iron I will have: first node - 2xXeon5607, 24GbRAM, 4x2Tb (I am planning raid10); second node - 2xXeon5404, 24GbRAM, 4x2Tb (I plan raid10); Storeydzh - XeonE31275, 8GbRAM, 16x2Tb (I plan raid10). I found howto to build a cluster failover with the KVM hypervisor, the example uses... Read more

Useful Links