Use 5 second isolation test timeout (#3907)

Sometimes isolation tests get stuck in CI and we cannot see why, because
the job is killed by the CI runner. This will instead fail inside make
the testsuite continue, but mark it as a failure like this in the diff
output:
```diff
+isolationtester: canceling step s2-ddl-create-index-concurrently after 5 seconds
 step s2-ddl-create-index-concurrently: CREATE INDEX CONCURRENTLY select_append_index ON select_append(id);
+ERROR:  CONCURRENTLY-enabled index command failed
```

We should detect blockages very quickly and the queries we run are also
very fast, so 5 seconds should be more than enough to catch any random
slowness. The default from Postgres is 5 minutes, which is waaay to much
for us.
pull/3849/head
Jelte Fennema 2020-06-16 14:57:49 +02:00 committed by GitHub
parent 799bfdab56
commit b71f82b31e
No known key found for this signature in database
GPG Key ID: 4AEE18F83AFDEB23
1 changed files with 4 additions and 0 deletions

View File

@ -15,6 +15,10 @@ endif
MAKEFILE_DIR := $(dir $(realpath $(firstword $(MAKEFILE_LIST))))
export PATH := $(MAKEFILE_DIR)/bin:$(PATH)
export PG_REGRESS_DIFF_OPTS = -dU10 -w
# Use lower isolation test timeout, the 5 minute default is waaay too long for
# us so we use 5 seconds instead. We should detect blockages very quickly and
# the queries we run are also very fast.
export PGISOLATIONTIMEOUT = 5
##
## Citus regression support