Add task retries smoke product test

[losipiuk]

Release notes

(x) This is not user-visible or is docs only, and no release notes are required.
( ) Release notes are required. Please propose a release note for me.
( ) Release notes are required, with the following suggested text:

# Section
* Fix some things. ({issue}`issuenumber`)

[martint]

Why does this need to be a product test instead of a regular test using the distributed query runner and multiple in-process nodes?

[losipiuk]

Why does this need to be a product test instead of a regular test using the distributed query runner and multiple in-process nodes?

Regular test did not catch jackson serialization problems introduced with #22722 (then reverted).
I do not fully understand the root cause - but classloader setup is different in DQR based tests than what we have in actual deployment - so I guess this is the problem.

DQR queries run fine but with product tests we get:

com.fasterxml.jackson.databind.exc.InvalidDefinitionException: No serializer found for class io.trino.plugin.base.metrics.TDigest and no properties discovered to create BeanSerializer (to avoid exception, disable SerializationFeature.FAIL_ON_EMPTY_BEANS) (through reference chain: io.tr
ino.execution.TaskInfo["stats"]->io.trino.operator.TaskStats["pipelines"]->com.google.common.collect.RegularImmutableList[0]->io.trino.operator.PipelineStats["operatorSummaries"]->com.google.common.collect.RegularImmutableList[0]->io.trino.operator.OperatorStats["pipelineMetrics"]->jav
a.util.ImmutableCollections$MapN["FileSystemExchangeSource.s3GetObjectRequestsFailed"]->io.trino.plugin.base.metrics.TDigestHistogram["digest"])
        at com.fasterxml.jackson.databind.exc.InvalidDefinitionException.from(InvalidDefinitionException.java:77)
        at com.fasterxml.jackson.databind.SerializerProvider.reportBadDefinition(SerializerProvider.java:1330)
        at com.fasterxml.jackson.databind.DatabindContext.reportBadDefinition(DatabindContext.java:414)
        at com.fasterxml.jackson.databind.ser.impl.UnknownSerializer.failForEmpty(UnknownSerializer.java:53)
        at com.fasterxml.jackson.databind.ser.impl.UnknownSerializer.serialize(UnknownSerializer.java:30)
        at com.fasterxml.jackson.databind.ser.BeanPropertyWriter.serializeAsField(BeanPropertyWriter.java:732)
        at com.fasterxml.jackson.databind.ser.std.BeanSerializerBase.serializeFields(BeanSerializerBase.java:770)
        at com.fasterxml.jackson.databind.ser.std.BeanSerializerBase.serializeWithType(BeanSerializerBase.java:653)
        at com.fasterxml.jackson.databind.ser.std.MapSerializer.serializeTypedFields(MapSerializer.java:1027)
        at com.fasterxml.jackson.databind.ser.std.MapSerializer.serializeFields(MapSerializer.java:779)
        at com.fasterxml.jackson.databind.ser.std.MapSerializer.serializeWithoutTypeInfo(MapSerializer.java:764)
        at com.fasterxml.jackson.databind.ser.std.MapSerializer.serialize(MapSerializer.java:720)
        at com.fasterxml.jackson.databind.ser.std.MapSerializer.serialize(MapSerializer.java:35)
        at com.fasterxml.jackson.databind.ser.std.JsonValueSerializer.serialize(JsonValueSerializer.java:289)
        at com.fasterxml.jackson.databind.ser.BeanPropertyWriter.serializeAsField(BeanPropertyWriter.java:732)
        at com.fasterxml.jackson.databind.ser.std.BeanSerializerBase.serializeFields(BeanSerializerBase.java:770)

on worker with #22722 applied.

[losipiuk]

Why does this need to be a product test instead of a regular test using the distributed query runner and multiple in-process nodes?

Regular test did not catch jackson serialization problems introduced with #22722 (then reverted). I do not fully understand the root cause - but classloader setup is different in DQR based tests than what we have in actual deployment - so I guess this is the problem.
....

#22797 fixes the problem - but it would be great to have regression test coverage.

[wendigo]

I don't think that adding a product test is worth it tbh. I'd prefer to create some test that basically creates a plugin class loader and tests whether you can serialize/deserialize class when you are in the server class loader.

[losipiuk]

I don't think that adding a product test is worth it tbh. I'd prefer to create some test that basically creates a plugin class loader and tests whether you can serialize/deserialize class when you are in the server class loader.

How would that be different from test which uses DistributedQueryRunner - where serialization works fine for some reason? This PT is extra 2 mins of CI runtime (in a job which is short anyway so does not extend overall CI for PR), so I do not think it is a big deal.

[wendigo]

Do we need to expose minio data to the host?