Upgrade of git.coopcloud.tech to 3.5.2+1.24.2-rootless fails #52

New Issue

2026-05-18T16:44:12Z

notplants commented

2026-05-18 16:44:12 +00:00

running into a failure that is causing a rollback during upgrade of git.coopcloud.tech:

$ abra app upgrade git.coopcloud.tech
? please select an upgrade (version: 3.4.0+1.24.2-rootless): 3.5.2+1.24.2-rootless
┏━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┓
┃                 UPGRADE OVERVIEW                 ┃
┃                                                  ┃
┃ DOMAIN                https://git.coopcloud.tech ┃
┃ RECIPE                gitea                      ┃
┃ SERVER                swarm-0.coopcloud.tech     ┃
┃ CONFIG                compose.yml                ┃
┃                       compose.mariadb.yml        ┃
┃                       compose.smtp.yml           ┃
┃                                                  ┃
┃ CURRENT DEPLOYMENT    3.4.0+1.24.2-rootless      ┃
┃ ENV VERSION           3.5.2+1.24.2-rootless      ┃
┃ NEW DEPLOYMENT        3.5.2+1.24.2-rootless      ┃
┗━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┛
WARN no release notes available for 3.5.2+1.24.2-rootless
? proceed? Yes
INFO initialising deployment
INFO polling deployment status
app: rolled back (retries: 0, healthcheck: healthy)
db: succeeded (retries: 0, healthcheck: ?)
FATA deploy failed 🛑
logs: /home/notplants/.abra/logs/swarm-0.coopcloud.tech/git_coopcloud_tech_2026-05-12T181804Z

the app seemed to e passing the healtcheck, but still moving to "rolled back"

running into a failure that is causing a rollback during upgrade of git.coopcloud.tech: ``` $ abra app upgrade git.coopcloud.tech ? please select an upgrade (version: 3.4.0+1.24.2-rootless): 3.5.2+1.24.2-rootless ┏━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┓ ┃ UPGRADE OVERVIEW ┃ ┃ ┃ ┃ DOMAIN https://git.coopcloud.tech ┃ ┃ RECIPE gitea ┃ ┃ SERVER swarm-0.coopcloud.tech ┃ ┃ CONFIG compose.yml ┃ ┃ compose.mariadb.yml ┃ ┃ compose.smtp.yml ┃ ┃ ┃ ┃ CURRENT DEPLOYMENT 3.4.0+1.24.2-rootless ┃ ┃ ENV VERSION 3.5.2+1.24.2-rootless ┃ ┃ NEW DEPLOYMENT 3.5.2+1.24.2-rootless ┃ ┗━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┛ WARN no release notes available for 3.5.2+1.24.2-rootless ? proceed? Yes INFO initialising deployment INFO polling deployment status app: rolled back (retries: 0, healthcheck: healthy) db: succeeded (retries: 0, healthcheck: ?) FATA deploy failed 🛑 logs: /home/notplants/.abra/logs/swarm-0.coopcloud.tech/git_coopcloud_tech_2026-05-12T181804Z ``` the app seemed to e passing the healtcheck, but still moving to "rolled back"

notplants referenced this issue

2026-05-18 16:44:43 +00:00

WIP: update some deprecated configs #51

notplants commented

2026-05-18 16:46:33 +00:00

I was able to find the container that exited and view its logs.

I think possibly healthcheck: healthy was just saying "healthy" for the old container... and there is some upgrade order of operations issues here?

I think the exited container (the new one), was failing to start completely and not even getting to the healthcheck. I see this in the logs:

2026/05/18 16:34:22 cmd/web.go:261:runWeb() [I] Starting Gitea on PID: 8
2026/05/18 16:34:22 cmd/web.go:114:showWebStartupMessage() [I] Gitea version: 1.24.2 built with GNU Make 4.4.1, go1.24.4 : bindata, timetzdata, sqlite, sqlite_unlock_notify
2026/05/18 16:34:22 cmd/web.go:115:showWebStartupMessage() [I] * RunMode: prod
2026/05/18 16:34:22 cmd/web.go:116:showWebStartupMessage() [I] * AppPath: /usr/local/bin/gitea
2026/05/18 16:34:22 cmd/web.go:117:showWebStartupMessage() [I] * WorkPath: /var/lib/gitea
2026/05/18 16:34:22 cmd/web.go:118:showWebStartupMessage() [I] * CustomPath: /var/lib/gitea/custom
2026/05/18 16:34:22 cmd/web.go:119:showWebStartupMessage() [I] * ConfigFile: /etc/gitea/app.ini
2026/05/18 16:34:22 cmd/web.go:120:showWebStartupMessage() [I] Prepare to run web server
2026/05/18 16:34:22 cmd/web.go:188:serveInstalled() [E] Unable to update WORK_PATH=/var/lib/gitea to config "/etc/gitea/app.ini": failed to save "/etc/gitea/app.ini": open /etc/gitea/app.ini: read-only file system
You should set it manually, otherwise there might be bugs when accessing the git repositories.
2026/05/18 16:34:22 routers/init.go:117:InitWebInstalled() [I] Git version: 2.49.0 (home: /var/lib/gitea/data/home)
2026/05/18 16:34:22 modules/setting/log.go:84:prepareLoggerConfig() [E] Deprecation: config option `[log].ENABLE_XORM_LOG` presents, please use `[log].logger.xorm.MODE` instead because this fallback will be/has been removed in 1.21
2026/05/18 16:34:22 modules/setting/mailer.go:79:loadMailerFrom() [E] Deprecation: config option `[mailer].MAILER_TYPE` presents, please use `[mailer].PROTOCOL` instead because this fallback will be/has been removed in v1.19.0
2026/05/18 16:34:32 modules/queue/manager.go:111:createWorkerPoolQueue() [E] Failed to create queue "mail": unable to lock level db at /var/lib/gitea/data/queues/common: resource temporarily unavailable
2026/05/18 16:34:32 services/mailer/mailer.go:61:NewContext() [F] Unable to create mail queue

I was able to find the container that exited and view its logs. I think possibly healthcheck: healthy was just saying "healthy" for the old container... and there is some upgrade order of operations issues here? I think the exited container (the new one), was failing to start completely and not even getting to the healthcheck. I see this in the logs: ``` 2026/05/18 16:34:22 cmd/web.go:261:runWeb() [I] Starting Gitea on PID: 8 2026/05/18 16:34:22 cmd/web.go:114:showWebStartupMessage() [I] Gitea version: 1.24.2 built with GNU Make 4.4.1, go1.24.4 : bindata, timetzdata, sqlite, sqlite_unlock_notify 2026/05/18 16:34:22 cmd/web.go:115:showWebStartupMessage() [I] * RunMode: prod 2026/05/18 16:34:22 cmd/web.go:116:showWebStartupMessage() [I] * AppPath: /usr/local/bin/gitea 2026/05/18 16:34:22 cmd/web.go:117:showWebStartupMessage() [I] * WorkPath: /var/lib/gitea 2026/05/18 16:34:22 cmd/web.go:118:showWebStartupMessage() [I] * CustomPath: /var/lib/gitea/custom 2026/05/18 16:34:22 cmd/web.go:119:showWebStartupMessage() [I] * ConfigFile: /etc/gitea/app.ini 2026/05/18 16:34:22 cmd/web.go:120:showWebStartupMessage() [I] Prepare to run web server 2026/05/18 16:34:22 cmd/web.go:188:serveInstalled() [E] Unable to update WORK_PATH=/var/lib/gitea to config "/etc/gitea/app.ini": failed to save "/etc/gitea/app.ini": open /etc/gitea/app.ini: read-only file system You should set it manually, otherwise there might be bugs when accessing the git repositories. 2026/05/18 16:34:22 routers/init.go:117:InitWebInstalled() [I] Git version: 2.49.0 (home: /var/lib/gitea/data/home) 2026/05/18 16:34:22 modules/setting/log.go:84:prepareLoggerConfig() [E] Deprecation: config option `[log].ENABLE_XORM_LOG` presents, please use `[log].logger.xorm.MODE` instead because this fallback will be/has been removed in 1.21 2026/05/18 16:34:22 modules/setting/mailer.go:79:loadMailerFrom() [E] Deprecation: config option `[mailer].MAILER_TYPE` presents, please use `[mailer].PROTOCOL` instead because this fallback will be/has been removed in v1.19.0 2026/05/18 16:34:32 modules/queue/manager.go:111:createWorkerPoolQueue() [E] Failed to create queue "mail": unable to lock level db at /var/lib/gitea/data/queues/common: resource temporarily unavailable 2026/05/18 16:34:32 services/mailer/mailer.go:61:NewContext() [F] Unable to create mail queue ```

notplants referenced this issue

2026-05-18 16:57:57 +00:00

change deploy logic to stop-first instead of start-first #53

Sign in to join this conversation.

1 Participants

Notifications

Due Date

No due date set.

Dependencies

No dependencies set.

Reference: coop-cloud/gitea#52