[jira] [Created] (IGNITE-8092) Put operation may hang if cache was destroyed asynchronously.

classic Classic list List threaded Threaded
1 message Options
Reply | Threaded
Open this post in threaded view
|

[jira] [Created] (IGNITE-8092) Put operation may hang if cache was destroyed asynchronously.

Anton Vinogradov (Jira)
Pavel Pereslegin created IGNITE-8092:
----------------------------------------

             Summary: Put operation may hang if cache was destroyed asynchronously.
                 Key: IGNITE-8092
                 URL: https://issues.apache.org/jira/browse/IGNITE-8092
             Project: Ignite
          Issue Type: Bug
          Components: cache
    Affects Versions: 2.4
            Reporter: Pavel Pereslegin


If there is more than one cache in the cache group then put operation on cache may hang if it was destroyed asynchronously.
For now this applies to all cache modes (PARTITIONED/REPLICATED/LOCAL) and to all atomicity modes (ATOMIC/TRANSACTIONAL).

This problem can not be reproduced if there is only one cache in the cache group.

Reproducer:
{code:java}
public class DestroyCacheTest extends GridCommonAbstractTest {

    private CacheConfiguration<Integer, Boolean> ccfg(String name, String grp) {
        return new CacheConfiguration<Integer, Boolean>(name).setCacheMode(CacheMode.PARTITIONED)
            .setAtomicityMode(CacheAtomicityMode.TRANSACTIONAL)
            .setWriteSynchronizationMode(CacheWriteSynchronizationMode.FULL_SYNC)
            .setGroupName(grp);
    }

    public void testDestroyAsync() throws Exception {
        String grpName = "testGroup";

        try (IgniteEx node = startGrid(0)) {
            node.createCache(ccfg("cache2", grpName));

            for (int n = 0; n < 100; n++) {
                IgniteCache<Integer, Boolean> cache1 = node.createCache(ccfg("cache1", grpName));

                AtomicInteger cntr = new AtomicInteger();

                GridTestUtils.runMultiThreadedAsync(() -> {
                    try {
                        int key;

                        while ((key = cntr.getAndIncrement()) < 10_000) {
                            if (key == 1000)
                                cache1.destroy();

                            cache1.putIfAbsent(key, true);
                        }
                    }
                    catch (Exception ignore) {
                        log.warning(ignore.getMessage());
                    }

                    return null;
                }, 6, "put-thread").get();
            }
        }
    }
}
{code}

p.s. for ATOMIC cache additional cache status check in GridCacheGateway#onStopped busy wait resolve this problem.




--
This message was sent by Atlassian JIRA
(v7.6.3#76005)