ignite-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Alexander Belyak (JIRA)" <j...@apache.org>
Subject [jira] [Created] (IGNITE-6758) Slow memory releasing while deactivation
Date Thu, 26 Oct 2017 04:53:00 GMT
Alexander Belyak created IGNITE-6758:

             Summary: Slow memory releasing while deactivation
                 Key: IGNITE-6758
                 URL: https://issues.apache.org/jira/browse/IGNITE-6758
             Project: Ignite
          Issue Type: Bug
      Security Level: Public (Viewable by anyone)
          Components: general
    Affects Versions: 2.1
            Reporter: Alexander Belyak
            Priority: Minor
             Fix For: 2.4

It take about 1 minutes to fill each page by 0 and release it to page pool in PageMemoryImpl.ClearSegmentRunnable()
from GridCacheDatabaseSharedManager.onCacheGroupsStopped(). When we have 100+M pages in hundred
of Gb of pageCache in take quite long to GridUnsafe.setMemory to 0 and in logs we get lot
of "Failed to wait for partition map exchange [topVer=AffinityTopologyVersion [topVer=56,
Ver=1], node=3676f020-0bf0-4145-861e-689c96d7e853]. Dumping pending objects that might be
the cause: "<no cause here> without any cause and progress indicator. So full grid reboot
take longer downtime with unnecessary warnings. 

This message was sent by Atlassian JIRA

View raw message