Show table extended is slow

[lipan1024]

Describe the problem
I use spark-thirft to connect to the spark cluster

When I use a console to run commands
dbt run -m test_spark5
This is the output log from the console

Found 331 models, 51 tests, 1 snapshot, 0 analyses, 577 macros, 0 operations, 12 seed files, 146 sources, 11 exposures

10:02:53 | Concurrency: 5 threads (target='dev')
10:02:53 |
10:02:53 | 1 of 1 START incremental model test.test_spark5...................... [RUN]
10:03:05 | 1 of 1 OK created incremental model test.test_spark5................. [OK in 12.67s]
10:03:06 |
10:03:06 | Finished running 1 incremental model in 53.37s.

When I run commands using both consoles
First console:dbt run -m test_spark5
Second console:dbt run -m test_spark6
This is the output log from the first console

Running with dbt=0.20.2
Found 331 models, 51 tests, 1 snapshot, 0 analyses, 577 macros, 0 operations, 12 seed files, 146 sources, 11 exposures

10:16:15 | Concurrency: 5 threads (target='dev')
10:16:15 |
10:16:15 | 1 of 1 START incremental model test.test_spark5...................... [RUN]
10:16:22 | 1 of 1 OK created incremental model test.test_spark5................. [OK in 6.60s]
10:16:22 |
10:16:22 | Finished running 1 incremental model in 111.36s.

This is the output log from the second console

Running with dbt=0.20.2
Found 331 models, 51 tests, 1 snapshot, 0 analyses, 577 macros, 0 operations, 12 seed files, 146 sources, 11 exposures

10:16:11 | Concurrency: 5 threads (target='dev')
10:16:11 |
10:16:11 | 1 of 1 START incremental model test.test_spark6...................... [RUN]
10:16:22 | 1 of 1 OK created incremental model test.test_spark6................. [OK in 10.33s]
10:16:22 |
10:16:22 | Finished running 1 incremental model in 110.50s.

Conclusion
The time for the same Spark Application to run the show table extended in test like '*' statement is doubled.

I have summed up this conclusion after testing for many times. Is there any way to avoid this problem.

Supplement
For some reason, I can only run one model at a time, so I can't use dbt run -m test_spark5 test_spark6

[jtcohen6]

(Relevant: dbt-labs/dbt-adapters#513)

[VersusFacit]

Hey there! First off @lipan1024 , were you able to discovery more or resolve your issue with:

For some reason, I can only run one model at a time, so I can't use dbt run -m test_spark5 test_spark6
If that's still an issue, it's probably worth a separate issue unless it was something related to your particular setup.

Second, the context you've given spells out the time difference well. That said, I can't quite tell why that show table instruction would differ. Some questions to help us hone in:

• Is this a regular occurrence?
• How did you isolate the show table instruction as the source of the performance issue?
• Do they have different schema they belong to? (this would affect the number of show table instructions; this may not be a good question according to certain things about your environment I can't know from afar)

[jtcohen6]

This is a known performance bottleneck, originally opened in dbt-labs/dbt-adapters#513. There's investigation and potential next steps described in #296