Jonathan,
> Is there any performance to be gained by having the workers pull
> files off of shared shared storage onto local storage, perform the
> workflows, and then putting the results back onto shared storage? Is
> this configuration even possible or desirable?
Yes. This is very useful if you have poor bandwidth but fast disks.
Reading/writing temp files over NFS can be a huge bottleneck.
> Also are you recommending different tiers of storage for the
> different settings? For example, should the workspace mounts be on a
> separate mount with faster storage?
We put everything over NFS. So our workers config looks like:
org.opencastproject.storage.dir=/data/opencast
org.opencastproject.file.repo.path=${org.opencastproject.storage.dir}/files
#org.opencastproject.file.repo.url=${org.opencastproject.admin.ui.url}
org.opencastproject.workspace.rootdir=${org.opencastproject.storage.dir}/workspace
org.opencastproject.workflow.solr.dir=${org.opencastproject.storage.dir}/workflow
So all of our workers share files over NFS. Our workspace directory
then looks something like:
http_mhadmin.usask.ca_8080
http_mhengage1.usask.ca_8080
http_mhworker2.usask.ca_8080
http_mhworker3.usask.ca_8080
http_mhworker4.usask.ca_8080
Since all of our machines share the workspace mount.
Having this on local storage would likely be faster, but since we're
using VMs for everything we don't have local storage options.
Chris
--
Christopher Brooks, BSc, MSc
ARIES Laboratory, University of Saskatchewan
Web: http://www.cs.usask.ca/~cab938
Phone: 1.306.966.1442
Mail: Advanced Research in Intelligent Educational Systems Laboratory
Department of Computer Science
University of Saskatchewan
176 Thorvaldson Building
110 Science Place
Saskatoon, SK
S7N 5C9
_______________________________________________
Matterhorn-users mailing list
[email protected]
http://lists.opencastproject.org/mailman/listinfo/matterhorn-users
