Efficient Analysis and Modernization of Computing Servers
Explore the rationale and goals behind the upgrade of computing servers at MEB in September 2016. The initiative aims to enhance efficiency, create an attractive computing environment, and ensure cost-effectiveness through resource sharing and scalability. Detailed technical specifications, requirements, and software/tools involved in the upgrade are outlined.
Download Presentation
Please find below an Image/Link to download the presentation.
The content on the website is provided AS IS for your information and personal use only. It may not be sold, licensed, or shared on other websites without obtaining consent from the author. If you encounter any issues during the download, it is possible that the publisher has removed the file from their server.
You are allowed to download the files provided on this website for personal or commercial use, subject to the condition that they are used lawfully. All files are the property of their respective owners.
The content on the website is provided AS IS for your information and personal use only. It may not be sold, licensed, or shared on other websites without obtaining consent from the author.
E N D
Presentation Transcript
Computing Servers at MEB 15/9/2016
Why? & Goals New Calculation Servers Architecture /Tech specs Old Vs New User Account File system Data transfer Guideline & Operation 15/9/2016
Why? Current environment is outdated. 15/9/2016
Goals Efficient Analysis - Achieve a more efficient approach to the users in which each user gets a good performance and capacity for their analysis Attractive Environment - Creating an attractive computing environment by offering a higher degree of parallelism and larger memory for computationally intensive tasks but not necessarily higher speed than standard PCs. Cost-efficiency - A cost-effective computing environment through resource sharing opportunity and scalability 15/9/2016
Requirements No Queuing system continue gentlemen s agreement. Storage Quota / Homefolders/Project Folders/CIFS Updated Operating System Secured Login (SAS) Updated applications Scratch disk 15/9/2016
Vector Matrix 15/9/2016
Technical Specifications Skjold IBM x3850 X5 4 x Intel Xeon E7- 8860 @ 2.27GHz Total 80 threads 256 GB RAM Vector Lenovo System x3850 X6 4 x Intel Xeon E7-8890 v3 @ 2.50GHz Total 144 threads 1,5 TB RAM Spjut Matrix IBM 3850 M2 2 x Intel Xeon X7350 @ 2.93GHz 8 threads 1 x IBM 785GB High IOPS MLC Mono Adapter Lenovo System x3650 M5 2 x Intel Xeon E5-2697 v3 @ 2.60GHz Total 56 threads 1 x Intel P3700 1,6TB NVMe Enterprise Performance Flash Adapter /saswork 768 GB RAM 100 GB RAM 15/9/2016
Vector Matrix SAS / Stata R, Plink, Gtool, gcta, snptest and more. 15/9/2016
Software/Tools Software updates periodically SAS R Stata Other applications(R, Plink, Gtool, gcta, snptest and more) New software policy and request on sharepoint. DO NOT send your software request to IT-Support directly. 15/9/2016
File system Home folder is on NFS. Same folder can be accessed from both the servers. Home folder quota - 100 GB per user Project folders are shared on both the servers. 15/9/2016
File system One home folder. Restricted personal storage space. Shared between two servers. Z: and P: can be accessed via /cifs/Z or /cifs/P Do not work directly against /cifs Project folders stay intact. /projects/<projectname> 15/9/2016
File system SAS and Stata work SAS work area and Stata temp area on Matrix server is on fast local storage. 150 GB max per user. Scratch Disk (Matrix/Vector) A temporary work space for all users. Use this space instead of home folders for fast temporary storage. /scratch/tmp/<username> 500 GB per user Warning: Files older than 5 days are automatically deleted. 15/9/2016
Data Transfer Users are responsible for moving the homefolder data from Spjut/Skjold to the new environment. Spjut / Skjold has their own home folders. Vector and Matrix share the same home folder. See intra.meb.ki.se for more information 15/9/2016
Data Transfer Delete unnecessary data. Check the data before/after the transfer. Data transfer via Kitty + WinSCP or use SCP directly from server to server. Delete the data on the old server when you are sure you moved everything you need. 15/9/2016
Old Vs New Spjut / Skjold Matrix/Vector - seperate Home folders /home/<username> - shared homefolder /nfs/home/<username> - User Quota on homefolder Limited (100 GB) - /scratch/tmp/<username> Scratch disk (500 GB) 15/9/2016
User accounts All current users are requested to send their request for account via intra.meb.ki.se and not via email to IT support. Request account ONLY on the server you use. All future users will have to go via a gatekeeper, who will guide you regarding the account on the server, depending on the type of analysis. 15/9/2016
Timeline/Operation 15/9/2016
Stata example A simple test was conducted, running the same Stata/IC 14.0 job under three environments (desktop, VDI, skjold). Time (in seconds) it took to run under each system: Desktop (Stata/IC 14): 2876 (48 minutes) VDI (Stata/IC 14): 5166 (86 minutes) Skjold (Stata/IC 14): 7463 (124 minutes) Skjold (Stata/MP 14): 3061 (51 minutes) 15/9/2016
Make use of your local computer if you can, It probably has i5 or i7 processor but you need more than 4GB RAM MEB IT provides better local machine to users if needed CPU i5 32 GB RAM Just send a request if you need more RAM to it-support@meb.ki.se 15/9/2016
Which is faster?? Local Computer Server 15/9/2016
Parallelism Efficient Resources 15/9/2016
Guidelines You can find the guideline documents on intra.meb.ki.se 15/9/2016
