Seafile Backups with restic #65
Labels
No Label
Breaking
Domain
komu.boo
Domain
langrock.info
Domain
libre.moe
Involves
Documentation
Involves
Security
Involves
Testing
Kind
Bug
Kind
Enhancement
Priority
Critical
Priority
High
Priority
Low
Priority
Medium
Service
Element
Service
Gitea
Service
Matrix
Service
Nextcloud
Service
Seafile
Service
Szuru|Kirei
Status
Abandoned
Status
Blocked
Status
Duplicate
Status
Invalid
Status
Need More Info
Status
Planned
Status
Won't Fix
No Milestone
No project
No Assignees
1 Participants
Notifications
Due Date
No due date set.
Dependencies
No dependencies set.
Reference: KomuSolutions/igot99issues#65
Loading…
Reference in New Issue
Block a user
No description provided.
Delete Branch "%!s()"
Deleting a branch is permanent. Although the deleted branch may continue to exist for a short time before it actually gets removed, it CANNOT be undone in most cases. Continue?
Currently using Scaleway Glacier for backups of Seafile.
With the recent change to restic for backups (#59) i wanted to also use it for Seafile backups.
Due to the offline-nature I cannot use restic with Scaleway Glacier, which is why I have been looking to change our backup target while staying cost-effective. Storj DCS costs 5,96€/TB (4$/TB + ~2,2$ segment fees per used TB), Backblaze 6,87€/TB (6$/TB + USt.) and GCP Archive 1,38€/TB (1,16€/TB + USt.) so I think the winner is clear. Only downside with GCP Archive-tier is their 1 year minimum retention, but when I configure restic to keep backups for a long time, this should not incur much cost - and the storage is quite cheap at that so I don't feel like keeping deleted data for long is a big issue - especially since restic is also encrypted.
My local HDD + tape backups are enough to plausibly restore all data in case of an error with primary storage, so it's unlikely I'll have to restore from GCP.
Seafile Backupsto Seafile Backups with resticDiscovered road block: restic in it's current configuration uses too much memory (as our data set consists of a huge amount of small files) but the host system does not have sufficient resources to allow restic such usage.
Since I run my setup off of juicefs, i could literally use any other server, mount the juciefs there and backup it up. Maybe I could find a service which allows me to run a job (like a container image set up for everything) and execute it with enough RAM/CPU/network)
If I used Scaleway for such backups, the costs would likely be extremly low for just running the backup daily and a clean up job every month or so.
lukas referenced this issue2024-12-27 14:53:44 +01:00
Closing for now, this setup would be complicated and left root credentials on another prodivers server. Mounting the fusefs might also proove difficult as I needed to somehow forward access to my redis db.
The local backups are versioned either way, so I could just make a new version there.