[jira] [Created] (IGNITE-6688) ignite.destroyCache method hangs the cluster

["Ender Ekici (JIRA)" <ji...@apache.org>]

Ender Ekici created IGNITE-6688:
-----------------------------------

             Summary: ignite.destroyCache method hangs the cluster
                 Key: IGNITE-6688
                 URL: https://issues.apache.org/jira/browse/IGNITE-6688
             Project: Ignite
          Issue Type: Bug
      Security Level: Public (Viewable by anyone)
    Affects Versions: 2.1
            Reporter: Ender Ekici


I've posted Apache Ignite Users but couldn't get an answer. [Thread|http://apache-ignite-users.70518.x6.nabble.com/ignite-destroyCache-method-hangs-the-cluster-tt16850.html]

Hi,
We have a test machine which has 8 cores with 32Gb ram.
On ignite there is only one cache which has 10M records (db size 7Gb~ ignite cache size 14Gb~)
We reload this cache with a cron scheduler.(Every night)

For example:
cacheName: products (alias)
realCacheName: products+oldTimestamp

-cron triggered.
-create cache with name “products+newTimestamp”
-load data from db to cache
-when loading finishes, we map the products cache to “products+newTimestamp” and later(after 1 min) destroy the old cache;

While we measuring the cache.get() times with only one client we encountered 4.5sec delay while ignite.destroyCache(realOldCacheName) operation.

Example Reproducer Code:
[Github|https://github.com/enderekici/ignitetest.git]



--
This message was sent by Atlassian JIRA
(v6.4.14#64029)