The most common connectivity problems stem from

• being off-campus but not connecting to the HBS VPN first (VPN instructions)
• being on-campus on the WiFi, but connected to HBS Guest instead of HBS Secure

Please see Logging In for more information.

Our NoMachine NX GUI login can be very particular at times. You may receive "connection errors", "access denied errors", or the like. A few simple steps may remedy the problem:

1. As noted in the other access FAQ, ensure you are connected via VPN if not on the HBS campus or on guest WiFi.
2. Try connecting via the terminal using the instructions on this page. If not successful, ensure that you're using the correct username / password combination. Otherwise, contact RCS.
3. If successful in connecting via terminal, ensure that you have completely exited NoMachine, and issue the command clean-nx-homedir in the terminal window. Now try connecting via NoMachine again.
4. As a last attempt, quit NoMachine. Remove the .nx directory present in your home folder on your Mac or PC, and try connecting via NoMachine again.

If you still cannot get in, please contact RCS.

There could be a number of reasons why your jobs are still sitting in the queue and haven't dispatched:

• See the FAQ: "I launched an application in NoMachine or Terminal. Why isn't it running?"
• See the FAQ: "How can I tell if the cluster is busy?"

If you've looked at the job details (bjobs -l JOBID) for the reason and your jobs do not start after a number of hours, contact RCS.

Sometimes you want to know the status of the cluster and scheduling -- this can help you determine why your jobs might be pending, or know that you need to reshape your jobs to ask for smaller resources to ensure you can get your jobs in and scheduled sooner than others.

Several commands, all in the terminal, will help give you a good picture:

• bqueues will show you the different queues, how many jobs are running, and how many are pending
• bjobs -u all will show you the list of jobs running and pending, and the order in which the scheduler is going to dispatch them. Note that this list/order will change frequently as resources become free and jobs finish sooner than expected
• bhosts will show all the nodes in the cluster, include max and used core counts
• lsload will also show all the nodes, but will also give %CPU utilization and well as how much memory is available for jobs

You can monitor a job either in Terminal / Command-Line via bjobs or by viewing the listing in the 'Jobs' window in Pac  (details here).

There are several reasons why an application may not appear to be running:

• All of the licenses for the application may be in use by others.
• More commonly, you may have reached the resource limit on your account (# of CPUs, memory limit, or # of interactive sessions).
• If the compute grid is busy, there many not be enough RAM or CPUs on any of the compute nodes to run your application.

Open a terminal session and type bjobs on the command line (details here). Look for line where the STATus is PEND. Type bjobs -l JOBID (replace JOBID with the ID of the job that is PENDing. Look for the PENDING REASONS section to understand why your job is not running.

See Troubleshooting LSF Jobs for more information.

With Grid 3.0, you can now ask for more RAM in an interactive session. See our documentation for more information.

Please see our documentation here.

JupyterLab is available on Grid 3.0. To get started with Grid 3.0, please see our documentation here.

This page demonstrates how you can check to see which version of R or RStudio is currently the default on Grid 3.0.

This page demonstrates how you can check to see which version of Python is currently the default on Grid 3.0.

If you accidentally deleted a file, wrote over code, or have to restore a file, please email RCS with the following information:

• Your name
• Full path of the file’s location
• File name
• Date file was deleted (if you do not know the exact date, a best guess is helpful)

We urge individuals to review files and folders to remove old, unused, or temporary work. As a part of this process, you can compress and archive groups of files. See our HOWTO document for this. If space is still needed, contact RCS to request more space.

This occurs when either you or a collaborator have not correctly set permissions. Subsequently, the owner of the file will need to change the file's permissions. You can identify the owner (and other permissions details) via the command ls -l MyFile.txt. Please see our documentaation on File Ownership and Permissions. If you need assistance from RCS, please include the output from ls -l MyFile.txt in your email.

File transfer programs, including the Mac Finder and Windows Explorer, try to do the best job possible matching permission when files are copies over. But since the operating systems are typically different between the source (your computer) and the destination (the storage on the compute grid), the programs often don't get it right. Please see our documentation here.