Cloudman - Startup Options - Galaxy Cluster Vs Data Cluster

Question: Cloudman - Startup Options - Galaxy Cluster Vs Data Cluster

6.9 years ago by

mailing list • 840 wrote:

Hi guys, When I first go to my cloudman page at <public dns="">/cloud I get a dialog asking for some settings. What does the Galaxy Cluster choice do? What does the Data Cluster choice do? Why can't I choose both? How much space should I allocate? Thanks, Greg Text of dialog: Galaxy Cluster: Galaxy application, available tools, reference datasets, SGE job manager, and a data volume. Specify the initial storage size (in Gigabytes): GBOK Share-an-Instance Cluster: derive your cluster form someone else's cluster. Specify the provided cluster share-string (for example, cm-0011923649e9271f17c4f83ba6846db0/shared/2011-08-19--21-00): Cluster share-string Data Cluster: a persistent data volume and SGE. Specify the initial storage size (in Gigabytes): GB Test Cluster: SGE only. No persistent storage is created.

galaxy • 1.0k views

ADD COMMENT • link •

modified 6.9 years ago by Dannon Baker ♦ 3.7k • written 6.9 years ago by mailing list • 840

6.9 years ago by

Dannon Baker ♦ 3.7k

United States

Dannon Baker ♦ 3.7k wrote:

Data Cluster only configures a persistent data volume and SGE. Galaxy Cluster includes a running Galaxy instance in addition to the persistent data volume and SGE configuration components from the Data Cluster option. The Galaxy Cluster option is what most users will probably want to choose. How much space is absolutely dependent on your needs. Keep in mind that you can increase size later if necessary via the admin UI. The maximum volume size (per an amazon limit to EBS volume size) is currently 1TB. -Dannon

ADD COMMENT • link written 6.9 years ago by Dannon Baker ♦ 3.7k

Thanks. So if I just want to run my own program that uses SGE, it sounds like I shouldn't need a Galaxy instance. Is that right? Does the storage space parameter control how much space is at /mnt/galaxyTools? -Greg

ADD REPLY • link written 6.9 years ago by mailing list • 840

It's also odd, if I launch CloudMan from the web interface here: https://biocloudcentral.herokuapp.com I don't see that dialog when I log into <public dns="">/cloud. Is it handling that aspect for me?

ADD REPLY • link written 6.9 years ago by mailing list • 840

Correct to the first point. The space available to /mnt/galaxyTools is not configurable through the interface, though you could manually resize a volume if you did need more room. /mnt/galaxyData is the volume referred to in the dialogs you're seeing. -Dannon

ADD REPLY • link written 6.9 years ago by Dannon Baker ♦ 3.7k

Hi guys, Another related question. So if I need a large data set for my program to run on, should that also be stored in GalaxyData? What is the best way to get the data on there? I saw something about FTP but it seemed kind of confusing. Also I chose the data cluster option so I don't think I have an instance of Galaxy running. Is that what would do the FTP? Thanks again, Greg

ADD REPLY • link written 6.9 years ago by mailing list • 840

Yes, galaxyData is the only persistent volume you have in your data only cluster, so everything should go there. SCP should work for getting data there, unless the file is already hosted somewhere in which case wget is probably a better choice. And yes, FTP is configured specifically for Galaxy uploads and is unavailable in the case of a non-galaxy cluster. The usernames used for logging in, etc, come straight from the Galaxy database.

ADD REPLY • link written 6.9 years ago by Dannon Baker ♦ 3.7k

Ok, yeah, SCP does make sense. Thanks. -Greg

ADD REPLY • link written 6.9 years ago by mailing list • 840

Similar posts • Search »