Task #7071

VRE #6301: VRE for ICES training course (ICES MSY): 25-26 Jan. 2017

Task #6938: Reinforce RStudio support to make it scalable and robust

Deploy RStudio Wrapper portlet to production

Added by Massimiliano Assante over 2 years ago. Updated over 2 years ago.

Status:ClosedStart date:Feb 10, 2017
Priority:UrgentDue date:Feb 13, 2017
Assignee:Massimiliano Assante% Done:

100%

Category:Default
Sprint:RStudio elastic management
Infrastructure:
Milestones:
Duration: 2

Related issues

Blocked by BlueBRIDGE - Task #7070: Please move the yet created 3 VMs hosting Rstudio in /gCu... Closed Feb 10, 2017 Feb 13, 2017

History

#1 Updated by Massimiliano Assante over 2 years ago

  • Blocked by Task #7070: Please move the yet created 3 VMs hosting Rstudio in /gCube/devNext/NextNext to production added

#2 Updated by Massimiliano Assante over 2 years ago

  • Status changed from New to In Progress

#3 Updated by Massimiliano Assante over 2 years ago

  • Status changed from In Progress to Feedback

#4 Updated by Massimiliano Assante over 2 years ago

  • % Done changed from 0 to 70

The portlet has been deployed, however the RStudio allocation may cause problems in case the same user has multiple VREs with RStudio.

#5 Updated by Massimiliano Assante over 2 years ago

  • % Done changed from 70 to 100
  • Status changed from Feedback to Closed

The previous problem is fixed in production, I just hot deployed the RStudio Wrapper portlet containing the fix.

Now the RStudio instance allocation:

  • takes into account the number of instances available in the given context and assign the RStudio instance balancing the load.
  • keep the assignee instance for that context until this instance exists in the context. If a problem occurs with that instance a new instance is assigned by looking at the one available.
  • If no instances are available the default instance (rstudio.d4science.org) is assigned.
  • maintains retro-compatibility with the previous version (that did not take into account the scope), meaning that if user foo had rstudio2 instance set on scope X, then in scope X foo is still assigned to rstudio2 (if valid)

Also available in: Atom PDF