@scooke eg. a use case is an open source community that uses a cloudron instance with some apps. They offer the apps to their community members. To have a consistence user management, it makes sense to have it in the ldap of Cloudron instead of in the user management of each individual application. If you only have 10 or 20 members it's not a problem to delegate the user creation to a user with the admin role. For all other numbers an open registration fits.

@jdaviescoates good questions

I have no experience with the Hetzner Cloud. So my answer is more or less a generic answer to a backup strategie.

I use snapshots provided by the provider for disaster recovery. If everything is broken, I can restore the complete image. Maybe the Hetzner backup service is a kind of "snapshot" (a solution to restore the complete server including all data).

I use the Cloudron backup (important for me: to a different storage (not the local files system) - mostly minio / S3) to backup all individual apps & the configuration of my Cloudron instance. This gives me the possibility

• clone my apps (in case of testing an update on a cloned app-instance before I update the app in a productive environment)
• restore all or some apps to another server, perhaps in a different hosting environment.

For your topic "over thinking": do you have a backup strategie for your backups on a backup server?

So in short:
For me, a backup from my provider is a solution for disaster recovery of the complete Cloudron instance in the infrastructure of the hoster.
Cloudron backups give me the possibility to restore some apps, clone them or switch to another hosting provider.

@jdaviescoates Yes. It is possible to connect both simultaneously. The usability in that case is not intuitive. But hey: you can

Since this version, OpenProject displays a banner for administrator accounts "Future versions of OpenProject will likely drop or reduce support for MySQL and MariaDB databases. Your installation is still running on MySQL and we suggest you migrate your installation to PostgreSQL. This process is easy when following our migration guides: Migrating your installation to PostgreSQL"
Do we have a roadmap for this change? I've heard from some guys from drupal, that they need PostgreSQL because of migration tasks.

@jdaviescoates good questions

I have no experience with the Hetzner Cloud. So my answer is more or less a generic answer to a backup strategie.

I use snapshots provided by the provider for disaster recovery. If everything is broken, I can restore the complete image. Maybe the Hetzner backup service is a kind of "snapshot" (a solution to restore the complete server including all data).

I use the Cloudron backup (important for me: to a different storage (not the local files system) - mostly minio / S3) to backup all individual apps & the configuration of my Cloudron instance. This gives me the possibility

• clone my apps (in case of testing an update on a cloned app-instance before I update the app in a productive environment)
• restore all or some apps to another server, perhaps in a different hosting environment.

For your topic "over thinking": do you have a backup strategie for your backups on a backup server?

So in short:
For me, a backup from my provider is a solution for disaster recovery of the complete Cloudron instance in the infrastructure of the hoster.
Cloudron backups give me the possibility to restore some apps, clone them or switch to another hosting provider.

@jdaviescoates Yes. It is possible to connect both simultaneously. The usability in that case is not intuitive. But hey: you can

@scooke eg. a use case is an open source community that uses a cloudron instance with some apps. They offer the apps to their community members. To have a consistence user management, it makes sense to have it in the ldap of Cloudron instead of in the user management of each individual application. If you only have 10 or 20 members it's not a problem to delegate the user creation to a user with the admin role. For all other numbers an open registration fits.

@girish yep. minio backend

@jdaviescoates Good idea. But please decide what OpenProject is. There is a community code base and some features are for "enterprise customers". Same with Mattermost. Ah. Matomo also have some paid options. Any suggestions?

Hi,

since I switched my backups from tgz to rsync, i got cloudron mails on a daily base.

Dear cloudron Admin,

Cloudron failed to create a complete backup. Please see https://cloudron.io/documentation/troubleshooting/#backups
for troubleshooting.

Logs for this failure are available at https://my.example.com/logs.html?taskId=0000

Old backup not found: snapshot/nextcloud/data/user/files/user/user/.~lock.name-03.odt#

My questions are:

• Does it means, that the backup is complete, but without the .~lock.name-03.odt file? Or does it stops after that error message?
• What is the problem with the "Old backup not found:" meaning? If there is no "old backup", put the file in the new one or trash it. Why is it a failure?
• For me as a human, my feeling about the .~lock.name-03.odt file is "forget about it. It's not important". That brings me to the question of a kind of blacklisting file types?!

Best
Stephan

Problem solved. With the oldest support tip: "Have you turned it off and on again?"
Conclusion: It was a local computer error.

Since the last update, no document in Nextcloud can be opened in ONLYOFFICE with Firefox. Chromium has no problems.
Maybe this? ->

Dec 05 11:33:03 [2019-12-05T11:33:03.820] [ERROR] nodeJS - changesError: docId = 1013443019 Error: TypeError: qe.zEb is not a function Script: https://onlyoffice.example.com/sdkjs/common/libfont/wasm/fonts.js Line: 40:388 userAgent: Mozilla/5.0 (Macintosh; Intel Mac OS X 10.14; rv:70.0) Gecko/20100101 Firefox/70.0 platform: MacIntel isLoadFullApi: false isDocumentLoadComplete: false StackTrace: @https://onlyoffice.example.com/sdkjs/common/libfont/wasm/fonts.js:40:388
Dec 05 11:33:03 @https://onlyoffice.example.com/sdkjs/common/libfont/wasm/fonts.js:40:395

My workaround: delete the content of the local.yaml.
https://github.com/kevinpapst/kimai2/issues/1290

@nebulon no manual editing or plugin installation. it seems that the instance has had this error for a while. But the good news: nope. It's not your update