• Task
  • Status: Closed
  • 0 Showstopper
  • Resolution: Fixed
  • ehcache,ehcache-monitor
  • hsingh
  • Reporter: amiller
  • November 06, 2009
  • 0
  • Watchers: 0
  • January 17, 2013
  • November 15, 2009

Description

Need to make sure that long-running test of the monitor doesn’t have memory leaks. Should test probes coming and going, caches coming and going, query functionality (if possible), etc.

Comments

Himadri Singh 2009-11-06

Test plan for ehcache-dx LRT:

  1. Start ehcache-dx server
  2. Start client with a dummy cacheManager registered to ehcache-dx server
  3. Add new cache managers to the clients. (There is no need to creating a new client to introduce new probes. Confirmed with Geert)
  4. Cache managers registered to ehcache-dx server should be random b/w 2-5
  5. Add cache to cache Manager. No. of caches should be random b/w 1-5
  6. Can add variety of replication modes. Non-replicated, Terracotta and RMI.
  7. Do some operations on the cache (gets, puts, etc) for a configurable duration.
  8. Get stats for the operations.
  9. Validate no. of cache managers/caches registered. Since gets and puts are occurring, validate the txn/sec doesn’t go below 5.
  10. Remove cache after operations
  11. Shutdown cache manager
  12. The duration should also be random, as don’t want all caches/cache manager registered/deregister at same time. (optional)
  13. LRT should repeat steps 3-12 after some interval.

Test properties

  1. Max cache managers
  2. Max caches to cache manager
  3. Duration of get-stats
  4. Test Interval

Himadri Singh 2009-11-15

Added another mode too, which starts server few min later than clients. Thus, clients are blocked waiting for server to get up.