[CI] XPackRestIT "analytics/top_metrics/sort by string script fails" failure

[astefan]

https://elasticsearch-ci.elastic.co/job/elastic+elasticsearch+7.x+multijob-unix-compatibility/os=oraclelinux-7&&immutable/574/console

02:31:08 org.elasticsearch.xpack.test.rest.XPackRestIT > test {p0=analytics/top_metrics/sort by string script fails} FAILED
02:31:08     java.lang.AssertionError: Failure at [analytics/top_metrics:247]: expected [400] status code but api [search] returned [200 OK] [{"took":8,"timed_out":false,"_shards":{"total":2,"successful":1,"skipped":0,"failed":1,"failures":[{"shard":0,"index":"test","node":"gtl65pvDSVSvSoO1viU9Lg","reason":{"type":"illegal_argument_exception","reason":"error building sort for [_script]: script sorting only supported on [numeric] scripts but was [string]"}}]},"hits":{"total":{"value":0,"relation":"eq"},"max_score":null,"hits":[]},"aggregations":{"tm":{"top":[]}}}]
02:31:08         at __randomizedtesting.SeedInfo.seed([6ECA51725C63A9D5:E69E6EA8F29FC42D]:0)
02:31:08         at org.elasticsearch.test.rest.yaml.ESClientYamlSuiteTestCase.executeSection(ESClientYamlSuiteTestCase.java:405)
02:31:08         at org.elasticsearch.test.rest.yaml.ESClientYamlSuiteTestCase.test(ESClientYamlSuiteTestCase.java:382)
02:31:08         at sun.reflect.GeneratedMethodAccessor14.invoke(Unknown Source)

Couldn't repro, though :-(.

[elasticmachine]

Pinging @elastic/es-analytics-geo (:Analytics/Aggregations)

[henningandersen]

Another failure in master here:
https://gradle-enterprise.elastic.co/s/muy4t6spxrake

Looks like the index somehow got created with 2 shards.

[henningandersen]

This master failure looks similar, but is for another test case: p0=analytics/top_metrics/sort by keyword field fails
https://gradle-enterprise.elastic.co/s/6expjc2t4e6rs/tests/fcijb24gqtmnc-q63yzuzm4b2wi
Looks like the index also got two shards in this case:

        "successful" : 1,
        "skipped" : 0,
        "failed" : 1,

[henningandersen]

Notice that an index template "global" is randomly added, causing the indices to have two shards.

@nik9000 I noticed this is a new test, maybe you can take a look? skip features: default_shards might do the trick here?

[nik9000]

Fixed in #52516.

[nik9000]

I had my wires crossed. This wasn't fixed.

[markharwood]

I might just have hit this in a PR build based off master abeb837 which has the number of shards fix. Didn't reproduce locally.

[nik9000]

I might just have hit this in a PR build based off master abeb837 which has the number of shards fix. Didn't reproduce locally.

I'll push some even more paranoid clamping on the test to see if that fixes it....

[nik9000]

More paranoia: 0a4c5e0