Discussion
1
Replies
290
Views
Incessant Technologies
Posted: 5 years ago
Last activity: 5 years 3 months ago
Managing High Available Applications : Q & A
Hi was going through the Managing High Available Applications Q & A.
Which component in a high availability architecture is useful to recover a single user session data in crash recovery? (Choose One)
Load balancer
Application Server Cache
Clustering model
Shared Storage
For this question , as per the Q& A correct answer is mention as Shared Storage.
But the question talks about single user session data in crash recovery. Single User session data recovery is required only when the browser crash. When the server Crashes, multiple user session will be involved. As per my understanding, For browser crash, application server cache is the better approach . Kindly suggest
Posted: 5 years ago
Hello Naveen,
See documentation below that may clarify the Q&A. Quote below is from page 36 of the document:
https://pdn.pega.com/documents/high-availability-administration-guide
Developing Highly Available Applications
Developing highly available applications and integrations in PRPC requires that System Architects and developers are aware of the High Available feature set, the PRPC prerequisites, and the general principles of high availability operation.
■ Quiescing a PRPC server to perform maintenance uses PRPC’s activation and passivation to shared storage. Users are passivated on quiescing PRPC servers and redirected at their next interaction to a new PRPC server in the cluster where their session is activated. Since quiesce relies on passivation and activation it works for all applications under all circumstances.
■ Crash recovery uses a new feature in Pega 7 that saves UI structure and metadata about work objects to shared storage. When a load balancer detects a PRPC server crash and removes that PRPC server from the pool of available servers, then users are directed to a new PRPC server. PRPC detects this crash condition and rebuilds the UI based on the UI structure and metadata.
■ Performing zero downtime PRPC upgrades uses split schema, with rules and data separated into different schemas. The process for performing a PRPC upgrade relies on split schema and PRPC’s quiesce capabilities to perform a rolling restart of all PRPC servers. Refer to the “split schema” and “Scenario: Performing a rolling restart” sections.
Lu