summaryrefslogtreecommitdiff
path: root/tools/test-setup-docker.sh
diff options
context:
space:
mode:
authorClark Boylan <clark.boylan@gmail.com>2023-01-30 13:41:15 -0800
committerClark Boylan <clark.boylan@gmail.com>2023-04-05 08:39:34 -0700
commit1c188ad440e82bb356832497c387aace71653dc8 (patch)
tree59b8cfcd36802b3a9438d7609f354f21fa962a77 /tools/test-setup-docker.sh
parentb2dc863b44d6b546f6609cfe8707f40c55b8aede (diff)
downloadzuul-1c188ad440e82bb356832497c387aace71653dc8.tar.gz
Test with MariaDB instead of MySQL
I'm not sure if this is a good idea or not yet. Testing with MySQL actually exposed the alembic transaction thing that is fixed in a grandparent change. That said much of the world seems to be moving to MariaDB so it might be a good idea for us to test that more explicitly? We update bindep.txt which will install mariadb during CI job runs. We also update our local test setup tooling to roughly match what will be used in CI. For example the mariadb:10.6 docker image is chosen because this is roughly equivalent to what Jammy packages. Another good reason for this change is that Debian ARM64 does not package MySQL and only packages MariaDB. This should give us better compatibility for testing on architectures other than x86-64. Change-Id: I9d00557ca5823da34278c3e9ecda2daaf13440b7
Diffstat (limited to 'tools/test-setup-docker.sh')
-rwxr-xr-xtools/test-setup-docker.sh2
1 files changed, 1 insertions, 1 deletions
diff --git a/tools/test-setup-docker.sh b/tools/test-setup-docker.sh
index 1601b11a7..6000230bb 100755
--- a/tools/test-setup-docker.sh
+++ b/tools/test-setup-docker.sh
@@ -38,7 +38,7 @@ else
fi
-MYSQL="${DOCKER} exec zuul-test-mysql mysql -u root -pinsecure_worker"
+MYSQL="${DOCKER} exec zuul-test-mariadb mysql -u root -pinsecure_worker"
if [ "${COMPOSE}" == "docker-compose" ]; then
${ROOTCMD} docker-compose rm -sf