Internal error checking workspace agent authorization (after db reinit from dump)

Hi. Today we tried to update Coder from v2.9.1 to v2.9.3 but the DB got corrupted (probably because of 2 pgsql instances running at the same time, rookie mistake). We have PG dumps each day, so I tried to run v2.9.3 version and import the backup dump (schema + data) from last known working version; the Coder works, I can access the dashboard, the data is there, but now all workspaces created show me the following error: Agent side:
2024-04-22 09:33:49.911 [info] connecting to coderd
2024-04-22 09:33:49.934 [warn] run exited with error ...
error= GET https://coder.company.com/api/v2/workspaceagents/me/rpc?version=2.0: unexpected status code 500: Internal error checking workspace agent authorization.
Error: pq: column "workspace_agents.created_at" must appear in the GROUP BY clause or be used in an aggregate function
2024-04-22 09:33:49.911 [info] connecting to coderd
2024-04-22 09:33:49.934 [warn] run exited with error ...
error= GET https://coder.company.com/api/v2/workspaceagents/me/rpc?version=2.0: unexpected status code 500: Internal error checking workspace agent authorization.
Error: pq: column "workspace_agents.created_at" must appear in the GROUP BY clause or be used in an aggregate function
Coder side:
2024-04-22 09:34:10.007 [warn] coderd: GET host=coder.company.com path=/api/v2/workspaceagents/me/rpc proto=HTTP/1.1 remote_addr=10.128.1.7 start="2024-04-22T09:34:10.006571993Z" took=883.371µs status_code=500 latency_ms=0 response_body="{\"message\":\"Internal error checking workspace agent authorization.\",\"detail\":\"pq: column \\\"workspace_agents.created_at\\\" must appear in the GROUP BY clause or be used in an aggregate function\"}\n" request_id=d9c36ff4-82a1-4a67-ae6a-c18ce26ddab5
2024-04-22 09:34:10.007 [warn] coderd: GET host=coder.company.com path=/api/v2/workspaceagents/me/rpc proto=HTTP/1.1 remote_addr=10.128.1.7 start="2024-04-22T09:34:10.006571993Z" took=883.371µs status_code=500 latency_ms=0 response_body="{\"message\":\"Internal error checking workspace agent authorization.\",\"detail\":\"pq: column \\\"workspace_agents.created_at\\\" must appear in the GROUP BY clause or be used in an aggregate function\"}\n" request_id=d9c36ff4-82a1-4a67-ae6a-c18ce26ddab5
I have tried investigating if the problem lies in some schema change, but that didn't help. I have also tried running v2.9.0, no change. If anybody has any idea, it would be very appreciated!
7 Replies
Codercord
Codercord6mo ago
<#1231901122399047690>
Category
Bug report
Product
Coder OSS (v2)
Platform
Linux
Logs
Please post any relevant logs/error messages.
modrin
modrin6mo ago
More detailed query:
2024-04-22 09:54:00.818 GMT [183] ERROR: column "workspace_agents.created_at" must appear in the GROUP BY clause or be used in an aggregate function at character 81

2024-04-22 09:54:00.818 GMT [183] STATEMENT: -- name: GetWorkspaceAgentAndOwnerByAuthToken :one

SELECT

workspace_agents.id, workspace_agents.created_at, workspace_agents.updated_at, workspace_agents.name, workspace_agents.first_connected_at, workspace_agents.last_connected_at, workspace_agents.disconnected_at, workspace_agents.resource_id, workspace_agents.auth_token, workspace_agents.auth_instance_id, workspace_agents.architecture, workspace_agents.environment_variables, workspace_agents.operating_system, workspace_agents.instance_metadata, workspace_agents.resource_metadata, workspace_agents.directory, workspace_agents.version, workspace_agents.last_connected_replica_id, workspace_agents.connection_timeout_seconds, workspace_agents.troubleshooting_url, workspace_agents.motd_file, workspace_agents.lifecycle_state, workspace_agents.expanded_directory, workspace_agents.logs_length, workspace_agents.logs_overflowed, workspace_agents.started_at, workspace_agents.ready_at, workspace_agents.subsystems, workspace_agents.display_apps, workspace_agents.api_version, workspace_agents.display_order,

workspaces.id AS workspace_id,

users.id AS owner_id,

users.username AS owner_name,

users.status AS owner_status,

workspaces.template_id AS template_id,

workspace_builds.template_version_id AS template_version_id,
2024-04-22 09:54:00.818 GMT [183] ERROR: column "workspace_agents.created_at" must appear in the GROUP BY clause or be used in an aggregate function at character 81

2024-04-22 09:54:00.818 GMT [183] STATEMENT: -- name: GetWorkspaceAgentAndOwnerByAuthToken :one

SELECT

workspace_agents.id, workspace_agents.created_at, workspace_agents.updated_at, workspace_agents.name, workspace_agents.first_connected_at, workspace_agents.last_connected_at, workspace_agents.disconnected_at, workspace_agents.resource_id, workspace_agents.auth_token, workspace_agents.auth_instance_id, workspace_agents.architecture, workspace_agents.environment_variables, workspace_agents.operating_system, workspace_agents.instance_metadata, workspace_agents.resource_metadata, workspace_agents.directory, workspace_agents.version, workspace_agents.last_connected_replica_id, workspace_agents.connection_timeout_seconds, workspace_agents.troubleshooting_url, workspace_agents.motd_file, workspace_agents.lifecycle_state, workspace_agents.expanded_directory, workspace_agents.logs_length, workspace_agents.logs_overflowed, workspace_agents.started_at, workspace_agents.ready_at, workspace_agents.subsystems, workspace_agents.display_apps, workspace_agents.api_version, workspace_agents.display_order,

workspaces.id AS workspace_id,

users.id AS owner_id,

users.username AS owner_name,

users.status AS owner_status,

workspaces.template_id AS template_id,

workspace_builds.template_version_id AS template_version_id,
Seems like v2.10.1 fixed our problems; there might be some internal schema problem between v2.9 and v2.10, or just a coincidence 🤷‍♂️
Cian
Cian6mo ago
We didn't make any schema changes from 2.9.0 -> 2.9.3 apart from a fix to remove a hardcoded public schema in a migration (https://github.com/coder/coder/issues/12619). That relates to a different table altogether and shouldn't actually change the resulting database schema. I tried reproducing your issue by running a fresh v2.9.1 install, upgrading to v2.9.3, and downgrading to v2.9.0 but couldn't reproduce the error you provided above. However, now that you're on v2.10 I would advise against downgrading.
modrin
modrin6mo ago
We're running coder from the start of v2 and it sounds like it might have something to do with existing PG dump data not importing correctly on a fresh v2.9 install. But for now the issue is fine on v2.10
Cian
Cian6mo ago
hmm.
modrin
modrin6mo ago
Maybe it's random 🤷‍♂️ idk if it happens again I'll report with more info 😄 Can't exactly share the dump with company data 😅 Would love to cleanup the dump (about 65 MB) sometime as well, some old remnants (after v0 to v2 migration) might be interfering with updates 🤔
Cian
Cian6mo ago
You can dump just the schema like we do here: https://github.com/coder/coder/blob/main/coderd/database/gen/dump/main.go#L53 I'm working on some extra tests around migration between versions right now, which is why this is of particular interest to me
Want results from more Discord servers?
Add your server