Weekly_2nd 6th_Oct 2017_DSpace.Status Report - Paulmulonzia/Dspace-test GitHub Wiki
DSpace Deployment in AWS
Weekly Status Report
For week ending 6/10/2017
| Milestone | Planned Dates | Actual Dates | Comments |
|---|---|---|---|
| Category 5 and 6 | |||
| Task 1: Decoupling Components & Layers with AWS capabilities | 02-Oct-17 | 03-Oct-17 | Accomplished |
| Task 2:Components & features isolation PoCs and Pilot testing | 03-Oct-17 | 06-Oct-17 | In progress |
Week’s Plan:
Improve on Category 5 (Decoupling Components & Layers with AWS capabilities) and complete Category 6 (Components & features isolation PoCs and Pilot testing)
Category 5's Summary
Decoupling Components & Layers with AWS capabilities.
We have moved from t2.large to t2.xlarge after adding additional two(2) DSpace instances(accounts) to make a single EC2 server run five(5) dspace instances(accounts).
Accomplishments
Making use of the following AWS services:
Route 53. We have 5 Alias record sets pointing to internet facing ELB which has our EC2 servers
- http://dspace.dddke.net (1st DSpace running on Tomcat 1 - port 8080)
- http://dspace1.dddke.net (2nd DSpace running on Tomcat 2 - port 8081)
- http://dspace2.dddke.net (3rd DSpace running on Tomcat 3 - port 8082)
- http://dspace3.dddke.net (4th DSpace running on Tomcat 4 - port 8083)
- http://dspace4.dddke.net (5th DSpace running on Tomcat 5 - port 8084)
N/B:
Each of the DSpace(account) can be accessed via the 5 subdomains above using the following credentials:
Email: [email protected]
Each subdomain represents a different DSpace Instance.
ELB. We have an internet-facing ELB which routes traffic to our Tomcat EC2 servers.
EC2 AMI. We have Apache, Tomcat and DSpace setup and installed.
DSpace installation directory mounted on EFS (/efs) a/efs/dspace, /efs/dspace1, /efs/dspace2, /efs/dspace3, /efs/dspace4)
AutoScaling. The architecture is now configured to scale up or down according to two policies which check MemoryUtilization (a custom metric). These policies are:
- dspace_increase: executed when the alarm Increase Group Size is breached (MemoryUtilization >= 90 for 300 seconds). When this happens, an instance is added to the Autoscaling Group and waits for 300 seconds before allowing another scaling activity.
- dspace_decrease: executed when the alarm Decrease Group Size is breached (MemoryUtilization <= 80 for 300 seconds). When this happens, an instance is removed from the Autoscaling Group using the NewestInstance termination policy and waits for 300 seconds before allowing another scaling activity.
RDS. We are using RDS Postgres DB Instance for our databases (dspacedb, dspacedb1, dspacedb2, dspacedb3 and dspacedb4).
EFS. We have successfully implemented and tested this, we have mounted the 3 dspace folders onto the EFS (/dspace, /dspace1, /dspace2, /dspace3 and /dspace4). This allows sharing of resources amongst EC2 Tomcat instances.
Category 6's Summary
Components & features isolation PoCs and Pilot testing.
Tasks
Task 1: Testing using jmeter
Task 2: CloudWatch Monitoring
Task 3: Grafana (RDS and EFS)
Accomplishment
Successfully tested & monitored based on the attached report
Plans for next week
Continue with Category 6 and start Category 7