Tag Info

New answers tagged

0

A consistent ring hash is a classic solution to this sort of problem, at least algorithmically. Systems like Swift, memcached, Dynamo, Riak, have used this approach to address this problem. OpenStack Swift specifically is an object/blob store that uses a consistent ring hash for tracking and distribution of objects, containers, etc. You could use Swift ...


0

As work around the communication difficulties, changing cluster communication from the default Multicast to Unicast: < <cman expected_votes="1" two_node="1"/> --- > <cman expected_votes="1" two_node="1" transport="udpu"/>


3

Obviously you need two switches. With LAG you need to end on the same physical switch (there are proprietary ways of this working in a switch cluster) but with LACP you can terminate on multiple devices and they coordinate the link. http://en.wikipedia.org/wiki/Link_aggregation


0

Although I followed all instructions on the web which are identical more or less, but apparently there is one missing step. I'm not sure it is because of my system (including software or hardware specifications). Anyway based of the instruction I found on the web by Michael Schwartzkopff (Linxu Magazine), it is necessary to run the following on both machines ...


0

Does your second node have a profile, and a running server or nodeagent running on it? You need it in order to the DMGR to being able to locate it an open a SOAP connection, to federate it. Also remember to enter the correct SOAP port, as well as the user/password for the 2nd machine instance.


0

DFS and DFS-R is not a solution to implement high availability. If you want to have highly available network shares, implement a SMB failover. Find some more information at http://blogs.technet.com/b/clausjor/archive/2012/06/07/smb-transparent-failover-making-file-shares-continuously-available.aspx


2

It is not possible to run multiple PostgreSQL servers from the same data directory, even if all but one are read-only. Absolutely 100% unsupported. Cannot be done. Give up now. Somebody might one day add such a feature but it'd involve major changes to PostgreSQL, as Pg relies heavily on shared memory and signals for inter-process synchronization. Also, the ...


6

You have many options. Which ones you can use depend on how your website is coded. Tightly Coupled, Single State This web-site can only run one instance because of... reasons. Running two in parallel would be an extremely bad idea for some reason. It's pretty uncommon to be this kind of site. CAP Theorem: Consistency over the whole system is paramount, not ...


2

Mosix has a different purpose. It's meant for parallel systems for HPC (high performance computing), not as a high availability solution.


1

If you can't telnet to port 9000 on the master from the slaves(and the error is connection timed out), then it's most likely a firewall issue. Check that the ec2 instances are in the same ec2 security group, and that there's no iptables firewall rules active blocking the connections(you can check this by running iptables -vnxL).


1

The Hyper-V service on Server 2012/2012 R2 has no need of the GUI, and it is easily managed from a nearby 8.1 workstation with RSAT, PowerShell or whatever. (Indeed, the Hyper-V Server product is so stripped down it doesn't include the GUI at all, nor many other things.)


2

To figure out if Server Core (or pretty much any other software) is right for you, you must answer the following questions in the affirmative: Will all the required roles run correctly on Server Core? (Some programs require the GUI to run/properly administer.) Does my team understand how to properly administer Server Core, or are willing to learn in an ...


0

Reading through the documentation for the various types of Matlab clusters, no where does it talk about using Windows Failover Clustering. See http://www.mathworks.com/help/mdce/configure-cluster.html


0

As background, Windows Failover Clustering is definitely the wrong technology for what you want. Right there in the name, it says "failover". It does not do distributed or compute clustering. Full stop - you're barking up the wrong tree. You should read up on Matlab's own documentation. Always read the vendor's stuff. MJS is a Matlab software component. If ...


0

I'm not certain whether this can or cannot be done with Apache + mod_proxy. As far as alternatives to Apache, there are other commercial solutions that provide load balancing and Websocket support. Shaka Technologies' Ishlangu Load Balancer ADC supports transparent (zero-configuration) websockets. I believe F5 BIG-IP do as well.


0

The HTTP 400 errors might be indicative of IIS rejecting the request due to no binding that matches the host header. This would happen if you've bound the app to an IP address which exists on one server but not the other. Your cluster resource group needs an IP address resource, which 'floats' between the nodes along with the app itself. This is in addition ...



Top 50 recent answers are included