# Does Temporal upgrade DB schemas automatically? **URL:** \ **Category:** Community Support\ **Created:** [April 3, 2023, 7:28pm UTC](https://community.temporal.io/t/does-temporal-upgrade-db-schemas-automatically/7777 "2023-04-03T19:28:46Z")\ **Posts on this page:** 10\ **Page:** 1
**Author:** ![RichRM](https://sea2.discourse-cdn.com/flex016/user_avatar/community.temporal.io/richrm/32/3640_2.png) [@RichRM](https://community.temporal.io/u/RichRM)\ **Post date:** [April 3, 2023, 7:28pm UTC](https://community.temporal.io/t/does-temporal-upgrade-db-schemas-automatically/7777/1 "2023-04-03T19:28:46Z")
For example, going from temporalio/server:1.19.1 to 1.20.1 I’m not clear if I need to manually run “auto-setup” again when upgrading, or if it does DB updates automatically (if needed). I think because these are consecutive (1.19 → 1.20) we don’t need to do anything special here and can just switch out the container image tag - is that right? Many thanks Rich ---
**Author:** ![tihomir](https://sea2.discourse-cdn.com/flex016/user_avatar/community.temporal.io/tihomir/32/6580_2.png) [@tihomir](https://community.temporal.io/u/tihomir)\ **Post date:** [April 3, 2023, 7:50pm UTC](https://community.temporal.io/t/does-temporal-upgrade-db-schemas-automatically/7777/2 "2023-04-03T19:50:11Z")
The docker image you update to includes the latest version binary and schema files. After update image version you should only need to update the existing schema, see for example [here](https://github.com/temporalio/docker-builds/blob/main/docker/auto-setup.sh#L191). I believe if you are running auto-setup, the auto-setup.sh script should take care of that part. ---
**Author:** ![RichRM](https://sea2.discourse-cdn.com/flex016/user_avatar/community.temporal.io/richrm/32/3640_2.png) [@RichRM](https://community.temporal.io/u/RichRM)\ **Post date:** [April 3, 2023, 7:54pm UTC](https://community.temporal.io/t/does-temporal-upgrade-db-schemas-automatically/7777/3 "2023-04-03T19:54:04Z")
That’s what I’m confused about. I used the auto-setup images while provisioning the server. Now we’re in run mode I switched to the “plain” server docker images (i.e. temporalio/server:1.19.1) So if I just swapped to temporalio/server:1.20.1, it would not automatically run the schema updates? Should we use the auto-setup image full time instead? ---
**Author:** ![tihomir](https://sea2.discourse-cdn.com/flex016/user_avatar/community.temporal.io/tihomir/32/6580_2.png) [@tihomir](https://community.temporal.io/u/tihomir)\ **Post date:** [April 3, 2023, 7:56pm UTC](https://community.temporal.io/t/does-temporal-upgrade-db-schemas-automatically/7777/4 "2023-04-03T19:56:12Z")
> Should we use the auto-setup image full time instead? No, it’s just for testing / local development as it runs all services in same process, and you cannot scale that for a production env. You have to run the update schema commands using the db-specific cli tool. ---
**Author:** ![RichRM](https://sea2.discourse-cdn.com/flex016/user_avatar/community.temporal.io/richrm/32/3640_2.png) [@RichRM](https://community.temporal.io/u/RichRM)\ **Post date:** [April 3, 2023, 7:58pm UTC](https://community.temporal.io/t/does-temporal-upgrade-db-schemas-automatically/7777/5 "2023-04-03T19:58:01Z")
Ah OK, so then the high level process would be: 1. Run a new container on the target version 2. Exec into it and run the temporal-sql-tool command 3. Upgrade the production containers to the new image Is that right? ---
**Author:** ![tihomir](https://sea2.discourse-cdn.com/flex016/user_avatar/community.temporal.io/tihomir/32/6580_2.png) [@tihomir](https://community.temporal.io/u/tihomir)\ **Post date:** [April 3, 2023, 8:03pm UTC](https://community.temporal.io/t/does-temporal-upgrade-db-schemas-automatically/7777/6 "2023-04-03T20:03:17Z")
I think the sql-tools are included in [admin-tools image](https://github.com/temporalio/docker-builds/blob/main/admin-tools.Dockerfile) so you could run schema updates from that. There is more info on updating server in docs [here](https://docs.temporal.io/cluster-deployment-guide#upgrade-server) as well if that helps. ---
**Author:** ![RichRM](https://sea2.discourse-cdn.com/flex016/user_avatar/community.temporal.io/richrm/32/3640_2.png) [@RichRM](https://community.temporal.io/u/RichRM)\ **Post date:** [April 3, 2023, 8:30pm UTC](https://community.temporal.io/t/does-temporal-upgrade-db-schemas-automatically/7777/7 "2023-04-03T20:30:30Z")
I’m on Postgres 13, should I still use “v96”? > ---
**Author:** ![RichRM](https://sea2.discourse-cdn.com/flex016/user_avatar/community.temporal.io/richrm/32/3640_2.png) [@RichRM](https://community.temporal.io/u/RichRM)\ **Post date:** [April 3, 2023, 9:07pm UTC](https://community.temporal.io/t/does-temporal-upgrade-db-schemas-automatically/7777/8 "2023-04-03T21:07:02Z")
I tried this with 12 (on a test cluster) and all seemed fine ```auto ## Main Schema update temporal-sql-tool --plugin postgres12 --ep "${POSTGRES_SEEDS}" -u "${POSTGRES_USER}" -pw "${POSTGRES_PWD}" -p "${DB_PORT}" --db "${DBNAME}" update-schema -d "/etc/temporal/schema/postgresql/v12/temporal/versioned" ## Visibility schema update temporal-sql-tool --plugin postgres12 --ep "${POSTGRES_SEEDS}" -u "${POSTGRES_USER}" -pw "${POSTGRES_PWD}" -p "${DB_PORT}" --db "temporal_visibility" update-schema -d "/etc/temporal/schema/postgresql/v12/visibility/versioned" ``` ---
**Author:** ![tihomir](https://sea2.discourse-cdn.com/flex016/user_avatar/community.temporal.io/tihomir/32/6580_2.png) [@tihomir](https://community.temporal.io/u/tihomir)\ **Post date:** [April 3, 2023, 9:35pm UTC](https://community.temporal.io/t/does-temporal-upgrade-db-schemas-automatically/7777/9 "2023-04-03T21:35:32Z")
Do you set DB env var to “postgres12”? If so yes. This should mean “12+” ---
**Author:** ![RichRM](https://sea2.discourse-cdn.com/flex016/user_avatar/community.temporal.io/richrm/32/3640_2.png) [@RichRM](https://community.temporal.io/u/RichRM)\ **Post date:** [April 3, 2023, 9:46pm UTC](https://community.temporal.io/t/does-temporal-upgrade-db-schemas-automatically/7777/10 "2023-04-03T21:46:19Z")
I’ve not set that env var anywhere in production (should we?), I was just trying to extract the upgrade command we need to run from the auto-setup script 🙂