An RStudio Server session will time out after 120 minutes of no user input. This is to make sure there are forgotten, left-over processes running on the system. For R tasks running longer than this, the solution is to submit them via the scheduler.
R is available on Wynton HPC via a contributed environment module. It can be run interactively in the terminal via R on a development node, as explain on the how-to ‘Work with R’ page. To run R via the RStudio GUI, launch your personal RStudio Server as instructed below and access it via your local web browser. As explained, this requires running two separate SSH connections to the cluster: (i) one to launch RStudio Server, and (ii) one to connect to it.
Assuming you are already logged on to a development node, launch your personal RStudio Server instance as:
[alice@dev1 ~]$ module load CBI rstudio-server-controller
[alice@dev1 ~]$ rsc start
alice, your personal RStudio Server 2023.09.1-494 running R 4.3.2 is available on:
<http://127.0.0.1:20612>
Importantly, if you are running from a remote machine without direct access
to dev1, you need to set up SSH port forwarding first, which you can do by
running:
ssh -L 20612:dev1:20612 alice@log1.wynton.ucsf.edu
in a second terminal from your local computer.
Any R session started times out after being idle for 120 minutes.
WARNING: You now have 10 minutes, until 2023-11-15 17:06:50-08:00, to
connect and log in to the RStudio Server before everything times out.
Your one-time random password for RStudio Server is: y+IWo7rfl7Z7MRCPI3Z4
There are two things you should pay extra attention to here:
The one-time random password that was generated
The instructions how to log in to the cluster with SSH port forwarding
You will need both below.
On your local computer, log into the cluster in a second terminal
following the instruction above. Make sure to use your own username
and make sure to use the port number (e.g. 20612) that
was assigned to you.
{local}$ ssh -L 20612:dev1.wynton.ucsf.edu:20612 alice@log1.wynton.ucsf.edu
alice1@log1.wynton.ucsf.edu:s password: XXXXXXXXXXXXXXXXXXX
[alice@log1 ~]$
If you successfully completed the above two steps, and you made sure to use the correct port, then you should be able to open your personal RStudio Server in your local web browser by going to:
You will be presented with a ‘Sign in to RStudio’ web page where you need to enter:
alice)y+IWo7rfl7Z7MRCPI3Z4)After clicking ‘Sign In’, you should be redirected to the RStudio interface.
To terminate the RStudio Server, start by exiting R by typing quit() at the R prompt. Then press Ctrl-C in the terminal where you called rsc start. Alternatively, run rsc stop in another terminal, e.g. the second one used in Step 2.
Some users report that they stuck when they try to log in to RStudio.
After they enter their username and password, and click ‘Sign In’,
they get to a page “R is taking longer to start than usual” with a
spinner that never ends. The user is presented with three options
‘Reload’, ‘Safe Mode’, and ‘Terminate R’. Ideally, ‘Safe Mode’ or
‘Terminate R’ would solve the problem and let the user access the
RStudio GUI. Unfortunately, for some users, none of these options
help. Consecutive attempts to use rsc stop and rsc start fail for
same reasons.
As of 2023-12-04, it is not clear why and when this happens. The one workaround we have found is to wipe the user’s RStudio sessions. For this, we recommend to use:
$ rsc reset --which=sessions
This will create a local copy of your problematic RStudio setup in
file rstudio-config_<timestamp>.tar, and then, only then, remove the
actually settings. The next time you call rsc start, you should
start out with a fresh RStudio setup, and the login issue should be
gone.
If you get the following error when launching rsc start:
[alice@dev2 ~]$ rsc start
WARNING: Needs to SSH to dev1.wynton.ucsf.edu to check whether process 2132343
is still alive [/wynton/home/boblab/alice/.config/rsc/rserver.hostname:
21 bytes; 2024-09-28 15:56:13)]. If you don't have SSH key authentication set up,
you will be asked to enter your account password below. The current machine is
dev2.wynton.ucsf.edu
ERROR: Failed to check process PID 2132343 on dev1.wynton.ucsf.edu over SSH.
Reason was: ssh: connect to host dev1.wynton.ucsf.edu port 22: No route to host
the reason is that rsc start tries to protect against launching more
than one RStudio session at the same time on different machines. In
order to confirm that you already running another RStudio session on
another machine, it needs to access that machine via SSH, but if that
fails you get the above error.
To troubleshoot this, start by making sure you can SSH to dev1.wynton.ucsf.edu. (1) If you can login manually, do that and call
rsc stop there. This should resolve the above problem. (2) If you
cannot access the machine, it could be that you have exhausted your
CPU quota on that machine and it is very slow to respond. If you
suspect this is the case, see ‘Running out of memory’ below. It could
also be that the machine is not working or down, which is rare, but it
happens. If it is down, it’s most likely already discussed on our
Slack forum - please check there to confirm it is truly down. In
the rare case that the machine is really down, try to call rsc reset
and the retry with rsc start. If you still get the above error,
retry with rsc reset --full.
If you get an ‘R Session Error’ dialog saying:
The previous R session was abnormally terminated due to an unexpected crash.
You may have lost workspace data as a result of this crash.
RStudio may not have restored the previously active project as a precaution. You may switch back to it using the Projects menu.
[OK]
one reason is that you ran out of memory and R was terminated by the operating system. Note that each user is limited to 96 GiB of RAM on the development node. Trying to use more than that, will cause the operating system to kill the underlying R process. When this happens, RStudio will likely keep running, but your R session was lost and reset.