Uploaded image for project: 'JGroups'
  1. JGroups
  2. JGRP-2398

LOCAL_PING should remove local PingData on stop

    XMLWordPrintable

Details

    • Bug
    • Resolution: Not a Bug
    • Major
    • 4.1.8
    • 4.1.6
    • None

    Description

      LOCAL_PING does not remove the local node's PING_DATA from the discovery map on stop.

      This means when a cluster name is reused, e.g. because we create a new cluster for each method in a test class, the nodes in the new cluster first try to connect to the dead coordinators of the old clusters. With the default GMS.max_join_attempts=10, this can take a really long time:

      13:53:39,024 DEBUG (testng-Test:[]) [GMS] address=Test-NodeA-57106, cluster=org.infinispan.partitionhandling.Test, physical address=127.0.0.1:55320
13:53:39,026 TRACE (testng-Test:[]) [GMS] Test-NodeA-57106: discovery took 0 ms, members: 17 rsps (4 coords) [done]
13:53:39,026 DEBUG (testng-Test:[]) [GMS] Test-NodeA-57106: found multiple coords: [054b27e3-9308-65bf-de26-d6cdd4f6bdd3, 6d7f9fb2-f56b-1b88-7f9d-cdee430ac8ca, Test-NodeA-37104, Test-NodeA-8303]
13:53:39,026 DEBUG (testng-Test:[]) [GMS] Test-NodeA-57106: sending JOIN(Test-NodeA-57106) to 6d7f9fb2-f56b-1b88-7f9d-cdee430ac8ca
13:53:41,026 WARN  (testng-Test:[]) [GMS] Test-NodeA-57106: JOIN(Test-NodeA-57106) sent to 6d7f9fb2-f56b-1b88-7f9d-cdee430ac8ca timed out (after 2000 ms), on try 0
13:53:41,026 DEBUG (testng-Test:[]) [GMS] Test-NodeA-57106: sending JOIN(Test-NodeA-57106) to Test-NodeA-37104
13:53:43,026 WARN  (testng-Test:[]) [GMS] Test-NodeA-57106: JOIN(Test-NodeA-57106) sent to Test-NodeA-37104 timed out (after 2000 ms), on try 0
13:53:43,026 DEBUG (testng-Test:[]) [GMS] Test-NodeA-57106: sending JOIN(Test-NodeA-57106) to 054b27e3-9308-65bf-de26-d6cdd4f6bdd3
13:53:45,027 WARN  (testng-Test:[]) [GMS] Test-NodeA-57106: JOIN(Test-NodeA-57106) sent to 054b27e3-9308-65bf-de26-d6cdd4f6bdd3 timed out (after 2000 ms), on try 0
13:53:45,027 DEBUG (testng-Test:[]) [GMS] Test-NodeA-57106: sending JOIN(Test-NodeA-57106) to Test-NodeA-8303
13:53:47,027 WARN  (testng-Test:[]) [GMS] Test-NodeA-57106: JOIN(Test-NodeA-57106) sent to Test-NodeA-8303 timed out (after 2000 ms), on try 0
...
13:54:51,036 DEBUG (testng-Test:[]) [GMS] Test-NodeA-57106: found multiple coords: [054b27e3-9308-65bf-de26-d6cdd4f6bdd3, 6d7f9fb2-f56b-1b88-7f9d-cdee430ac8ca, 0eb381de-3e9a-0320-b2d9-5db3cb39bb34, Test-NodeA-8303]
13:54:51,036 DEBUG (testng-Test:[]) [GMS] Test-NodeA-57106: sending JOIN(Test-NodeA-57106) to 054b27e3-9308-65bf-de26-d6cdd4f6bdd3
13:54:53,036 WARN  (testng-Test:[]) [GMS] Test-NodeA-57106: JOIN(Test-NodeA-57106) sent to 054b27e3-9308-65bf-de26-d6cdd4f6bdd3 timed out (after 2000 ms), on try 9
13:54:53,036 DEBUG (testng-Test:[]) [GMS] Test-NodeA-57106: sending JOIN(Test-NodeA-57106) to 0eb381de-3e9a-0320-b2d9-5db3cb39bb34
13:54:55,037 WARN  (testng-Test:[]) [GMS] Test-NodeA-57106: JOIN(Test-NodeA-57106) sent to 0eb381de-3e9a-0320-b2d9-5db3cb39bb34 timed out (after 2000 ms), on try 9
13:54:55,037 DEBUG (testng-Test:[]) [GMS] Test-NodeA-57106: sending JOIN(Test-NodeA-57106) to 6d7f9fb2-f56b-1b88-7f9d-cdee430ac8ca
13:54:57,037 WARN  (testng-Test:[]) [GMS] Test-NodeA-57106: JOIN(Test-NodeA-57106) sent to 6d7f9fb2-f56b-1b88-7f9d-cdee430ac8ca timed out (after 2000 ms), on try 9
13:54:57,037 DEBUG (testng-Test:[]) [GMS] Test-NodeA-57106: sending JOIN(Test-NodeA-57106) to Test-NodeA-8303
13:54:59,037 WARN  (testng-Test:[]) [GMS] Test-NodeA-57106: JOIN(Test-NodeA-57106) sent to Test-NodeA-8303 timed out (after 2000 ms), on try 9
13:54:59,037 WARN  (testng-Test:[]) [GMS] Test-NodeA-57106: too many JOIN attempts (10): becoming singleton

      Attachments

        Issue Links

          relates to

          Bug - A problem that impairs or prevents the functions of the product. ISPN-10937 CleanupAfterMethod tests take a very long time

          • Major - To be worked after higher priority work items are addressed. When the severity is high and the effort to change it is low to moderate, this term is chosen. Issues in this category likely have an existing workaround but implementation or execution may be non-trivial.
          • Closed

          Activity

            People

              rhn-engineering-bban Bela Ban
              dberinde@redhat.com Dan Berindei (Inactive)
              Votes:
              0 Vote for this issue
              Watchers:
              2 Start watching this issue

              Dates

                Created:
                Updated:
                Resolved: