Cloudron Forum

Hi,

it would be nice to enable short-lived ReadWrite access for container images, especially for quick and dirty hacks and fixes.

For example: In the last Synapse update, there was a bug using version x of Pillow where a deprecated attribute (Image.ANTIALIAS) was gone. This made it impossible to upload images to Matrix. Changing a line of code or installing PIL in version y could solve the problem. But not in the ReadOnly container.

I would like to see a button "Enable RW in container until restart" for non-persistent changes/fixes/debugging/something else.

A short reminder that all changes are lost on restart would be helpful.

@potemkin_ai
full ack!

its like the thing a few years ago, when the etherpad major upgrade kills all my pluginfull etherpad instances.

it's a tricky situation. to apply no updates is bad. to apply all updates is risky and more or less in schedule not synced with my worktimes or holidays or other ressources.

all or nothing or something between. hard to decide...

I use a nfs target for backups. and sometimes i have to restart some servers or routers or vpn. and even sometimes after this the backup volume is not mounted.

then the next backup fails. and the only thing i have to do is to click "remount storage" one time, 1 second later the mountpoint is green again and then i can manually start a new backup or wait for the automatic one.

it would be nice, if the backup task clicks the remount storage button for me

i am not sure if this is a feature request or a bug report

Yesterday i had to restore my synapse server from backup. worked.

but: only the database was broken. the restore of the database runs some seconds and all is fine. after this i had to wait one and a half hour while the media data (files on filesystem) where restored. my problem: only the database was broken. to solve my primary problem, the one minute restore of the database has could been the solution.

i would prefer to select in a advanced tab, which parts of my backup should be restored. the same thing as with nextcloud or wordpress or other tools.

It would be a nice feature to make a decision if an app uses the central database or its own container. in my docker-compose stacks i mostly have single purpose database containers for services like synapse. it is a question of taste, but i like single purpose database containers more than centralized storage.

There are Apps, where it makes no sense to run a dedicated Database Server. But on Things like Synapse or Mastodon it makes sense to split it from other Apps. Even in case of need for performing tweaks.

@girish

thank you for the fast help! the system is up and running again!

@holm

the named app b1f26524-5ea6-43a0-bcd7-da77b20fb136 is a disabled gitlab instance.

Hi,

after upgrading to 7.4.1 there is the "Starting services, this can take a while" Message and all Database based Apps stucks.

systemctl restart box

brings in the log:

2023-04-17T15:45:30.750Z box:services pipeRequestToFile: connected with status code 404
2023-04-17T15:45:30.750Z box:services exportDatabase: Error exporting postgresql of app b1f26524-5ea6-43a0-bcd7-da77b20fb136. BoxError: Unexpected response code or HTTP error when piping http://172.18.30.2:3000/databases/dbb1f265245ea643a0bcd7da77b20fb136/backup?access_token=VERYLONGTOKEN to /home/yellowtent/appsdata/b1f26524-5ea6-43a0-bcd7-da77b20fb136/postgresqldump: status 404
    at ClientRequest.<anonymous> (/home/yellowtent/box/src/services.js:1224:60)
    at ClientRequest.emit (node:events:513:28)
    at HTTPParser.parserOnIncomingClient (node:_http_client:693:27)
    at HTTPParser.parserOnHeadersComplete (node:_http_common:128:17)
    at Socket.socketOnData (node:_http_client:534:22)
    at Socket.emit (node:events:513:28)
    at addChunk (node:internal/streams/readable:315:12)
    at readableAddChunk (node:internal/streams/readable:289:9)
    at Socket.Readable.push (node:internal/streams/readable:228:10)
    at TCP.onStreamRead (node:internal/stream_base_commons:190:23) {
  reason: 'Addons Error',
  details: {}
}
2023-04-17T15:45:30.751Z box:platform Failed to start services. retry=false (attempt 0): Unexpected response code or HTTP error when piping http://172.18.30.2:3000/databases/dbb1f265245ea643a0bcd7da77b20fb136/backup?access_token=VERYLONGTOKEN to /home/yellowtent/appsdata/b1f26524-5ea6-43a0-bcd7-da77b20fb136/postgresqldump: status 404
2023-04-17T15:45:30.751Z box:cloudron Startup task at index 2 failed: Unexpected response code or HTTP error when piping http://172.18.30.2:3000/databases/dbb1f265245ea643a0bcd7da77b20fb136/backup?access_token=VERYLONGTOKEN to /home/yellowtent/appsdata/b1f26524-5ea6-43a0-bcd7-da77b20fb136/postgresqldump: status 404 BoxError: Unexpected response code or HTTP error when piping http://172.18.30.2:3000/databases/dbb1f265245ea643a0bcd7da77b20fb136/backup?access_token=VERYLONGTOKEN to /home/yellowtent/appsdata/b1f26524-5ea6-43a0-bcd7-da77b20fb136/postgresqldump: status 404
    at ClientRequest.<anonymous> (/home/yellowtent/box/src/services.js:1224:60)
    at ClientRequest.emit (node:events:513:28)
    at HTTPParser.parserOnIncomingClient (node:_http_client:693:27)
    at HTTPParser.parserOnHeadersComplete (node:_http_common:128:17)
    at Socket.socketOnData (node:_http_client:534:22)
    at Socket.emit (node:events:513:28)
    at addChunk (node:internal/streams/readable:315:12)
    at readableAddChunk (node:internal/streams/readable:289:9)
    at Socket.Readable.push (node:internal/streams/readable:228:10)
    at TCP.onStreamRead (node:internal/stream_base_commons:190:23)

thats all i can see what ist curious. what to do next?

i think, that i have done all steps from https://forum.cloudron.io/topic/8169/starting-services-this-can-take-a-while-on-v7-3-4-update and https://forum.cloudron.io/topic/9042/apps-not-starting-after-upgrade-to-7-4-1 but t seams for me, that the problem is in postgresql / postgresql startup tasks.

@girish it would be great if this behavior was described in the cloudron package documentation. it's good that i have now fixed all my email problems in cloudron, but it would have been even better if i had known that the ü in my last name was causing the problems i wanted to work around