Vertx cluster wide map - entry is not removed after ttl expires

[Mirko Stojmenovic]

Hi,

I'm using Vert.x 4.1.1, and I'm having issue with entry not being removed from cluster wide map after ttl expires. I wrote a reproducer unit test, that passes on Vert.x 3.5.4, but fails on 4.1.1. Do you have any idea why the entry is not removed from map after ttl expires?

Reproducer:

    @Test
    public void testEntryTtl(TestContext context) {
        Async async = context.async();

        Vertx.clusteredVertx(new VertxOptions(), clusteredVertxResult -> {
                         if (clusteredVertxResult.failed()) {
                             context.fail(clusteredVertxResult.cause());
                         }
                         Vertx vertx = clusteredVertxResult.result();

                         String mapName = "mapName";

                         String key = "key";
                         String value = "value";
                         long entryTtl = 200;

                         SharedData sharedData = vertx.sharedData();
                         sharedData.<String, String>getClusterWideMap(
                                 mapName, (createMap) -> {
                             if (createMap.failed()) {
                                 context.fail(createMap.cause());
                             }
                             AsyncMap<String, String> clusterSharedMap = createMap.result();
                             clusterSharedMap.put(key, value, entryTtl, (
                                                  putEntry) -> {
                                              if (putEntry.failed()) {
                                                  context.fail(putEntry.cause());
                                                  return;
                                              }
                                              vertx.setTimer(entryTtl + 10,
                                                             (timer) -> {
                                                         clusterSharedMap.get(key, (getEntry) -> {
                                                                          if (getEntry.failed()) {
                                                                              context.fail(getEntry.cause());
                                                                          }
                                                                          String result = getEntry.result();
                                                                          System.out.println("Result: " + result);
                                                                          context.assertNull(result);
                                                                          async.complete();
                                                                      });
                                                     });

                                          });
                         });
                     });
    }

[Julien Viet]

it looks like a regression.

could you open an issue along with the reproducer ?

Julien

> --
> You received this message because you are subscribed to the Google Groups "vert.x" group.
> To unsubscribe from this group and stop receiving emails from it, send an email to vertx+un...@googlegroups.com.
> To view this discussion on the web, visit https://groups.google.com/d/msgid/vertx/9097a8a2-c895-4a10-8c2d-35142d64c9e5n%40googlegroups.com.

[Thomas SEGISMONT]

Which cluster manager are you using ?

2021-07-29 15:59 UTC+02:00, Julien Viet <jul...@julienviet.com>:

> https://groups.google.com/d/msgid/vertx/CA%2BH-X4-XbR3Vdw2jSxTXB1uj9MsHmFgEvgNLnLT5ujEmZHwqEw%40mail.gmail.com.
>

[Mirko Stojmenovic]

I'm using Hazelcast cluster manager.

[Thomas SEGISMONT]

I tried this test with Vert.x 4.1.1.

It seems like the issue comes from a limitation of the corresponding  Hazelcast version (4.0.1) which rounds the ttl up to the nearest second.

If you upgrade to the latest Hazelcast version (4.2.1) the test passes.

I filed this issue: https://github.com/vert-x3/vertx-hazelcast/issues/145

To view this discussion on the web, visit https://groups.google.com/d/msgid/vertx/98da9b06-dbb5-4ea2-9f3a-671e758cbe5en%40googlegroups.com.

[Mirko Stojmenovic]

Perfect, thank you for confirming the issue, and thank you for the suggestion on how to fix it.

[Thomas SEGISMONT]

You're welcome.

To view this discussion on the web, visit https://groups.google.com/d/msgid/vertx/3e1205bf-2403-4640-9e2c-e9919549b54dn%40googlegroups.com.