WeBWorK Main Forum

Clustering and High Availability Webwork Servers

Clustering and High Availability Webwork Servers

by Danny Glin -
Number of replies: 2
Does anyone have any experience with any sort of clustering of webwork servers? We are currently using LVS with round robin scheduling to distribute the http traffic among two nodes, but we do not have a solution in place that allows us to distribute requests to the mysql server. We are looking for a solution which offers both distribution of load and redundancy of the sql server, so that there is no single point of failure. I am currently investigating the mysql cluster suite. Does anyone have experience with this or other high-availability solutions with regard to webwork?
In reply to Danny Glin

Re: Clustering and High Availability Webwork Servers

by Michael Gage -
Hi Danny,

I think you may be first here. We've tried to keep in mind the idea that the mysql database and the problem rendering operations might take place on machines other than the webserver itself, but we don't get enough load at the University of Rochester to make it a high priority experiment to work on. Let us know how it goes.

-- Mike


In reply to Danny Glin

Re: Clustering and High Availability Webwork Servers

by Christian McHugh -
We've got a clustered setup going now.
http://wwrk.maa.org/moodle/mod/forum/discuss.php?d=548

We're using heartbeat/LVS to load balance two webservers and two mysql servers. The mysql servers keep in sync through mysql replication (clustered mysql did not work as it would require schema changes to the webwork databases)