Question
Tech Mahindra Limited
IN
Posted: Sep 25, 2018
Last activity: Sep 26, 2018
Last activity: 26 Sep 2018 1:43 EDT
Closed
Rolling Restart of NBA server nodes on every 15 days/Once in the month for Application maintenance activities.
As a best practices and avoid following issues, we need to reboot all NBA server nodes one by one on every 15 days/ Once in the month for Application maintenance activities.. Could you please provide your view on this. Thanks!
Issue: Following issue will be occurred if we do not reboot server nodes at least 15 days/once in the month.
• Garbage collection is not happens properly hence lot of agents will be stuck
• Sync between server nodes are not happens properly
• Dnode joining failed issue
• Performance issue
• Memory leak issue
• Marketing and decision agents will be stuck
• System Pulse will be always high
• Lot of latency issues in NBA
• Requestor lock exceptions
• Long running sessions
• Wrong rule running and erratic system pulse, for example
System pulse is suspected of not keeping rules synchronized across nodes.
Pulse fails to pick up or apply a rule change on one or more nodes.
Show More
As a best practices and avoid following issues, we need to reboot all NBA server nodes one by one on every 15 days/ Once in the month for Application maintenance activities.. Could you please provide your view on this. Thanks!
Issue: Following issue will be occurred if we do not reboot server nodes at least 15 days/once in the month.
• Garbage collection is not happens properly hence lot of agents will be stuck
• Sync between server nodes are not happens properly
• Dnode joining failed issue
• Performance issue
• Memory leak issue
• Marketing and decision agents will be stuck
• System Pulse will be always high
• Lot of latency issues in NBA
• Requestor lock exceptions
• Long running sessions
• Wrong rule running and erratic system pulse, for example
System pulse is suspected of not keeping rules synchronized across nodes.
Pulse fails to pick up or apply a rule change on one or more nodes.
Rule resolution or rule cache issues appear in a multi-node deployment or on a single node (unknown causes related to system pulse).
• Hung threads
• Thread dumps that can cause system performance issues
• Out of memory errors
• Static content errors
• Failure to regenerate a Rule-Utility-Function (RUF) or Rule-Utility-Library (RUL) after a RAP import
• Looping rules running rampant
• Rule caching issues, for example, a higher version is available, but is not found (development server)
• Static cache issue causing the user interface (images, icons, other elements controlled by the CSS) to not display correctly
Benefits:
• Healthy and Stable NBA environment
• All functionality will be work smoothly
• Prevent Latency/slowness issue
• No data Sync Issues
• Garbage collection working properly
• No Memory leaks
• System Pulse will be always low
• All agents/Listeners will be run properly
• Dnode joining will be always Normal
• Increase performance in Marketing portal and it’s components
• Audience/segments will populate faster
Show Less
Posted: 6 years ago
Pegasystems Inc.
IN
Hi Satish,
It seems we are here looking for a process to be followed as best practice.
In case your environment is subjected to huge business load, it would be wise to adopt such practices of schedule restarts.
It does help cleaning up a lot of stale thread, memory resources etc.
Regards,
Ujjwal