RStudio Server Pro Power Tools

Customize Login Page

Customize Login Page

Edit:

/etc/rstudio/login.html

• Code is injected after the login box.
• Use CSS to position elements elsewhere on the page
• Use JS to modify the page

High Availability

• Use multiple RStudio Server Pro instances as failovers for each other
• Provide service to analysts even if one or more machines are down
• Supported using RStudio Server Pro “Multiple Masters” load balanced configuration

High Availability

Log everything that goes to the R console

session_id, project, pid, username, timestamp, type, data
125a3b4f, "", 28519, "rstudiouser", 1483482800226, prompt, "> "
125a3b4f, "", 28519, "rstudiouser", 1483482818498, input, "demo(package = .packages(all.available = TRUE))"
125a3b4f, "", 28519, "rstudiouser", 1483482818559, prompt, "> "
125a3b4f, "", 28519, "rstudiouser", 1483482826169, input, "demo(nlm)"
125a3b4f, "", 28519, "rstudiouser", 1483482826176, output, "\n\n        demo(nlm)\n     ---- ~~~\n"
125a3b4f, "", 28519, "rstudiouser", 1483482827311, input, ""
125a3b4f, "", 28519, "rstudiouser", 1483482827319, output, "\n> #  Copyright (C) 1997-2009 The R Core Team\n> \n> ### Helical Valley Function\n> ### Page 362 Dennis + Schnabel\n> \n> require(stats); require(graphics)\n\n> theta <- function(x1,x2) (atan(x2/x1) + (if(x1 <= 0) pi else 0))/ (2*pi)\n\n> ## but this is easier :\n> theta <- function(x1,x2) atan2(x2, x1)/(2*pi)\n\n> f <- function(x) {\n+     f1 <- 10*(x[3] - 10*theta(x[1],x[2]))\n+     f2 <- 10*(sqrt(x[1]^2+x[2]^2)-1)\n+     f3 <- x[3]\n+     return(f1^2+f2^2+f3^2)\n+ }\n\n> ## explore surface {at x3 = 0}\n> x <- seq(-1, 2, length.out=50)\n\n> y <- seq(-1, 1, length.out=50)\n\n> z <- apply(as.matrix(expand.grid(x, y)), 1, function(x) f(c(x, 0)))\n\n> contour(x, y, matrix(log10(z), 50, 50))\n"
125a3b4f, "", 28519, "rstudiouser", 1483482827443, prompt, "Hit <Return> to see next plot: "
125a3b4f, "", 28519, "rstudiouser", 1483482828392, input, ""
125a3b4f, "", 28519, "rstudiouser", 1483482828434, output, "\n> str(nlm.f <- nlm(f, c(-1,0,0), hessian = TRUE))\nList of 6\n $ minimum   : num 1.24e-14\n $ estimate  : num [1:3] 1.00 3.07e-09 -6.06e-09\n $ gradient  : num [1:3] -3.76e-07 3.49e-06 -2.20e-06\n $ hessian   : num [1:3, 1:3] 2.00e+02 -4.07e-02 9.77e-07 -4.07e-02 5.07e+02 ...\n $ code      : int 2\n $ iterations: int 27\n\n> points(rbind(nlm.f$estim[1:2]), col = ""red"", pch = 20)\n\n> ### the Rosenbrock banana valley function\n> \n> fR <- function(x)\n+ {\n+     x1 <- x[1]; x2 <- x[2]\n+     100*(x2 - x1*x1)^2 + (1-x1)^2\n+ }\n\n> ## explore surface\n> fx <- function(x)\n+ {   ## `vectorized' version of fR()\n+     x1 <- x[,1]; x2 <- x[,2]\n+     100*(x2 - x1*x1)^2 + (1-x1)^2\n+ }\n\n> x <- seq(-2, 2, length.out=100)\n\n> y <- seq(-0.5, 1.5, length.out=100)\n\n> z <- fx(expand.grid(x, y))\n\n> op <- par(mfrow = c(2,1), mar = 0.1 + c(3,3,0,0))\n\n> contour(x, y, matrix(log10(z), length(x)))\n"
125a3b4f, "", 28519, "rstudiouser", 1483482828549, prompt, "Hit <Return> to see next plot: "
125a3b4f, "", 28519, "rstudiouser", 1483482829218, input, ""
125a3b4f, "", 28519, "rstudiouser", 1483482829295, output, "\n> str(nlm.f2 <- nlm(fR, c(-1.2, 1), hessian = TRUE))\nList of 6\n $ minimum   : num 3.97e-12\n $ estimate  : num [1:2] 1 1\n $ gradient  : num [1:2] -6.54e-07 3.34e-07\n $ hessian   : num [1:2, 1:2] 802 -400 -400 200\n $ code      : int 1\n $ iterations: int 23\n\n> points(rbind(nlm.f2$estim[1:2]), col = ""red"", pch = 20)\n\n> ## Zoom in :\n> rect(0.9, 0.9, 1.1, 1.1, border = ""orange"", lwd = 2)\n\n> x <- y <- seq(0.9, 1.1, length.out=100)\n\n> z <- fx(expand.grid(x, y))\n\n> contour(x, y, matrix(log10(z), length(x)))\n\n> mtext(""zoomed in"");box(col = ""orange"")\n\n> points(rbind(nlm.f2$estim[1:2]), col = ""red"", pch = 20)\n\n> par(op)\n\n> fg <- function(x)\n+ {\n+     gr <- function(x1, x2) {\n+         c(-400*x1*(x2 - x1*x1)-2*(1-x1), 200*(x2 - x1*x1))\n+     }\n+     x1 <- x[1]; x2 <- x[2]\n+     res<- 100*(x2 - x1*x1)^2 + (1-x1)^2\n+     attr(res, ""gradient"") <- gr(x1, x2)\n+     return(res)\n+ }\n\n> nlm(fg, c(-1.2, 1), hessian = TRUE)\n$minimum\n[1] 1.182096e-20\n\n$estimate\n[1] 1 1\n\n$gradient\n[1]  2.583521e-09 -1.201128e-09\n\n$hessian\n        [,1]    [,2]\n[1,]  802.24 -400.02\n[2,] -400.02  200.00\n\n$code\n[1] 1\n\n$iterations\n[1] 24\n\n\n> ## or use deriv to find the derivatives\n> \n> fd <- deriv(~ 100*(x2 - x1*x1)^2 + (1-x1)^2, c(""x1"", ""x2""))\n\n> fdd <- function(x1, x2) {}\n\n> body(fdd) <- fd\n\n> nlm(function(x) fdd(x[1], x[2]), c(-1.2,1), hessian = TRUE)\n$minimum\n[1] 1.182096e-20\n\n$estimate\n[1] 1 1\n\n$gradient\n[1]  2.583521e-09 -1.201128e-09\n\n$hessian\n        [,1]    [,2]\n[1,]  802.24 -400.02\n[2,] -400.02  200.00\n\n$code\n[1] 1\n\n$iterations\n[1] 24\n\n\n> fgh <- function(x)\n+ {\n+     gr <- function(x1, x2)\n+         c(-400*x1*(x2 - x1*x1) - 2*(1-x1), 200*(x2 - x1*x1))\n+     h <- function(x1, x2) {\n+         a11 <- 2 - 400*x2 + 1200*x1*x1\n+         a21 <- -400*x1\n+         matrix(c(a11, a21, a21, 200), 2, 2)\n+     }\n+     x1 <- x[1]; x2 <- x[2]\n+     res<- 100*(x2 - x1*x1)^2 + (1-x1)^2\n+     attr(res, ""gradient"") <- gr(x1, x2)\n+     attr(res, ""hessian"") <- h(x1, x2)\n+     return(res)\n+ }\n\n> nlm(fgh, c(-1.2,1), hessian = TRUE)\n$minimum\n[1] 2.829175\n\n$estimate\n[1] -0.6786981  0.4711891\n\n$gradient\n[1] -0.4911201  2.1115987\n\n$hessian\n         [,1]     [,2]\n[1,] 366.1188 271.4593\n[2,] 271.4593 200.0000\n\n$code\n[1] 4\n\n$iterations\n[1] 100\n\n"
125a3b4f, "", 28519, "rstudiouser", 1483482829500, prompt, "> "

Log everything that goes to the R console

Set:

audit-r-console=all

in /etc/rstudio/rserver.conf. It’s also possible to log input only (specify input rather than all)

Try a new version of R

Try a new version of R

1. Compile new version of R
2. Put in a well-known location such as /opt/R, or add path to /etc/rstudio/r-versions
3. Restart RStudio Server Pro

Choose project when signing in

Choose project when signing in

Get notified when RSP goes down

Get notified when RSP goes down

1. Enable health check with server-health-check-enabled=1
2. Install infrastructure monitoring tool such as Nagios and point it to health check
3. Configure notifications

Customize session launch

1. Set environment variables
2. Configure access to clusters
3. Mount folders with data
4. Etc.

Customize session launch

Method 1: Modify

/etc/rstudio/rsession-profile

Method 2: Set

rsession-exec-command=your-script

Disable publishing and downloading

Disable publishing and downloading

Set these options in /etc/rstudio/rsession.conf:

allow-file-downloads=0
allow-external-publish=0

Give power users priority

Give power users priority

Set in /etc/rstudio/profiles:

[*]
cpu-affinity = 1-4
max-processes = 100
max-memory-mb = 2048
session-timeout-minutes=60

[@powerusers]
cpu-affinity = 5-16
nice = -10
max-memory-mb = 4096

Use a project-specific version of R

Use a project-specific version of R

Store new projects outside home dir

Store new projects outside home dir

In /etc/rstudio/rsession.conf:

session-default-new-project-dir=/mnt/bigdisk/projects

Also, consider setting R_USER, which allows customization of your “user home dir” as seen by R.

Stable mapping of users to load balanced nodes

Stable mapping of users to load balanced nodes

In /etc/rstudio/load-balancer

[config]
balancer = user-hash

Impersonate a user

Impersonate a user

admin-superuser-group=rstudio-super-users

Set default settings for new users

Set default settings for new users

1. Set the settings you want in a user account
2. Extract them from the user’s ~/.rstudio/monitored/user-settings folder
3. Copy to /etc/skel

Log R session activity

pid, username, timestamp, type, data
28450, "rstudiouser", 1483482767000, auth_login, ""
28519, "rstudiouser", 1483482769000, session_start, ""
28450, "rstudiouser", 1483484148000, auth_login, ""
28450, "rstudiouser", 1483484258000, auth_logout, ""
28450, "rstudiouser", 1483484265000, auth_login, ""
28450, "rstudiouser", 1483484277000, auth_logout, ""
28450, "rstudiouser", 1483484283000, auth_login_failed, ""
28450, "rstudiouser", 1483484287000, auth_login, ""
28519, "rstudiouser", 1483484297000, session_quit, ""
28519, "rstudiouser", 1483484297000, session_exit, ""
28796, "rstudiouser", 1483484308000, session_start, ""
28796, "rstudiouser", 1483484317000, session_quit, ""
28796, "rstudiouser", 1483484317000, session_exit, ""
28835, "rstudiouser", 1483484318000, session_start, ""
28835, "rstudiouser", 1483484321000, session_quit, ""
28835, "rstudiouser", 1483484321000, session_exit, ""
28871, "rstudiouser", 1483485057000, session_start, ""

Log R session activity

In /etc/rstudio/rserver.conf:

audit-r-sessions=1

Immortal sessions

Immortal sessions

In /etc/rstudio/profiles:

[*]
session-timeout-minutes=0

Can also designate specific groups or users who have immortal sessions.

Open project in a new session

Open project in a new session

See another session’s output

See another session’s output

1. Split the output stream from the source session with sink(file = "/tmp/foo", split = TRUE)
2. Open file as a source file in another session
3. See what your other session’s up in one click and without leaving your current session window

With some modifications, this technique is also useful for consuming partial output of long-running processes.

Graphite monitoring

Graphite monitoring

In /etc/rstudio/rserver.conf

monitor-graphite-enabled=1
monitor-graphite-host=134.47.22.6
monitor-graphite-port=2003

Pair program and code review

Pair program and code review

The End