Sapelo2 Frequently Asked Questions

From Research Computing Center Wiki
Revision as of 16:33, 12 December 2017 by Gcormier (talk | contribs)
Jump to navigation Jump to search

When can I get access to Sapelo2?

We are still configuring and testing Sapelo2 and it is not yet available for users. Sapelo users will be notified before we start migrating accounts from Sapelo to Sapelo2. If necessary, e.g. if you have jobs running on Sapelo, we will coordinate with you before migrating your account.

How do I connect to Sapelo2

Once we migrate your account from Sapelo to Sapelo2, you will be able to connect to Sapelo2 using Secure Shell (ssh). The hostname will be made available when Sapelo2 is ready for users. You will need to authenticate using your UGA MyID and you will need to use the two-factor (Duo) authentication.

When connecting to Sapelo2 from an off-campus location, users will need to use the UGA VPN.

For more information on how to sign up for Duo, please use your MyID to log into [1]

How to use the UGA VPN is detailed at [2]

Will I have to transfer data between Sapelo and Sapelo2?

No. Your home directory on Sapelo2 will be the same as on Sapelo. The Lustre file system will also be available on Sapelo2 and it will be mounted with the same path, i.e. /lustre1. Therefore, Sapelo users will not have to transfer data to Sapelo2.

Can I use the xfer nodes to transfer data to/from Sapelo2?

Yes. You can continue to use the file transfer nodes (xfer nodes) to transfer data between your local machine and Sapelo2 and to transfer data between your Sapelo2 home and /lustre1 directories, and to access data in your lab's project area (/project).

What operating system does Sapelo2 use?

The operating system (OS) on Sapelo2 is CentOS Linux release 7.1. Note that Sapelo uses CentOS Linux 6.5. Therefore, if you will use any applications installed in your home directory or code that you previously compiled, we'd suggest that you recompile and reinstall them on Sapelo2.

What queueing system does Sapelo2 run?

Sapelo2 uses the same queueing system that is used on Sapelo, namely Torque/Moab. However, it is likely that there will be some small configuration changes on Sapelo2 and you might have to make some (mostly minor) changes to your job submission scripts. One significant change is that users will be required to request memory for their jobs and memory usage will be enforced. We suggest that current Sapelo users monitor memory usage of their current jobs, if needed, to glean a better understanding of memory consumption by their different types of jobs.

What software packages are available on Sapelo2?

We are still working on installing applications on Sapelo2. There are over 500 software packages installed centrally on Sapelo and it is our plan to migrate those that users still need to Sapelo2. While we do not have resources to migrate all needed packages to Sapelo2 before its go-live date, we will continue to add packages after Sapelo2 is in production. Our intention is to install first those applications that are needed by most users or that are dependencies for other applications.

A preliminary list of applications that have already been installed on Sapelo2 is available at Software installed on Sapelo2.