roachtest: tpcdsvec failed

[cockroach-teamcity]

roachtest.tpcdsvec failed with artifacts on master @ 4f361d30d788f9614695e687b386aedd783ee3ba:

The test failed on branch=master, cloud=gce:
test artifacts and logs in: /artifacts/tpcdsvec/run_1
	tpcdsvec.go:175,tpcdsvec.go:184,test_runner.go:875: ERROR: query execution canceled due to statement timeout (SQLSTATE 57014)
		(1) secondary error attachment
		  | ERROR: query execution canceled due to statement timeout (SQLSTATE 57014)
		  | (1)
		  | Wraps: (2) ERROR: query execution canceled due to statement timeout (SQLSTATE 57014)
		  | Error types: (1) *fmt.wrapError (2) *pgconn.PgError
		Wraps: (2) secondary error attachment
		  | ERROR: query execution canceled due to statement timeout (SQLSTATE 57014)
		  | (1)
		  | Wraps: (2) ERROR: query execution canceled due to statement timeout (SQLSTATE 57014)
		  | Error types: (1) *fmt.wrapError (2) *pgconn.PgError
		Wraps: (3)
		Wraps: (4) ERROR: query execution canceled due to statement timeout (SQLSTATE 57014)
		Error types: (1) *secondary.withSecondaryError (2) *secondary.withSecondaryError (3) *fmt.wrapError (4) *pgconn.PgError

Help

See: roachtest README

See: How To Investigate (internal)

/cc @cockroachdb/sql-queries

This test on roachdash | Improve this report!

Jira issue: CRDB-13977

[rytaft]

The two queries that failed are:

06:54:44 tpcdsvec.go:141: test status: running query 7
07:04:44 tpcdsvec.go:149: test status: encountered an error: ERROR: query execution canceled due to statement timeout (SQLSTATE 57014)
vectorize=OFF:

SELECT
	i_item_id,
	avg(ss_quantity) AS agg1,
	avg(ss_list_price) AS agg2,
	avg(ss_coupon_amt) AS agg3,
	avg(ss_sales_price) AS agg4
FROM
	store_sales,
	customer_demographics,
	date_dim,
	item,
	promotion
WHERE
	ss_sold_date_sk = d_date_sk
	AND ss_item_sk = i_item_sk
	AND ss_cdemo_sk = cd_demo_sk
	AND ss_promo_sk = p_promo_sk
	AND cd_gender = 'F'
	AND cd_marital_status = 'W'
	AND cd_education_status = 'College'
	AND (p_channel_email = 'N' OR p_channel_event = 'N')
	AND d_year = 2001
GROUP BY
	i_item_id
ORDER BY
	i_item_id
LIMIT
	100;
;

vectorize=ON: [same as previous]

and

07:09:30 tpcdsvec.go:141: test status: running query 26
07:19:30 tpcdsvec.go:149: test status: encountered an error: ERROR: query execution canceled due to statement timeout (SQLSTATE 57014)
vectorize=OFF:

SELECT
	i_item_id,
	avg(cs_quantity) AS agg1,
	avg(cs_list_price) AS agg2,
	avg(cs_coupon_amt) AS agg3,
	avg(cs_sales_price) AS agg4
FROM
	catalog_sales,
	customer_demographics,
	date_dim,
	item,
	promotion
WHERE
	cs_sold_date_sk = d_date_sk
	AND cs_item_sk = i_item_sk
	AND cs_bill_cdemo_sk = cd_demo_sk
	AND cs_promo_sk = p_promo_sk
	AND cd_gender = 'M'
	AND cd_marital_status = 'W'
	AND cd_education_status = 'Unknown'
	AND (p_channel_email = 'N' OR p_channel_event = 'N')
	AND d_year = 2002
GROUP BY
	i_item_id
ORDER BY
	i_item_id
LIMIT
	100;
;

vectorize=ON: [same as previous]

[rytaft]

I don't think a timeout should be a release blocker. Leaving as GA blocker until we understand the reason for the timeout.

[yuzefovich]

There is a third query that timed out too.

The test uses 5 minutes for each query. Queries 7, 26, and 60 timed out. According to the numbers from 21.2 release time frame in #37464, all these queries should be much faster than 5 minutes, so it appears we have some regressions on master.

[yuzefovich]

Last night's run completed without timeouts, so I think it was the streamer bug to blame.