Delete nextcloud/data/*-old-ssd rollback copies (hdd-bulk migration) #37

Closed
opened 2026-07-19 03:03:59 +00:00 by claude-bot · 1 comment
Collaborator

Nextcloud's per-user data directories were migrated from swarm-data to hdd-bulk (USB HDD tier) on 2026-07-19, keeping the pre-migration copies as <user>-old-ssd on swarm-data as a rollback net (same pattern as the swarm-nvme migration, issue #6).

Delete once the hdd-bulk copies have proven stable for ~1-2 weeks (target: ~2026-08-02):

cd /mnt/swarm-data/nextcloud/data
rm -rf *-old-ssd

This is what actually reclaims the ~332GiB on swarm-data.

Nextcloud's per-user data directories were migrated from `swarm-data` to `hdd-bulk` (USB HDD tier) on 2026-07-19, keeping the pre-migration copies as `<user>-old-ssd` on `swarm-data` as a rollback net (same pattern as the swarm-nvme migration, issue #6). Delete once the hdd-bulk copies have proven stable for ~1-2 weeks (target: ~2026-08-02): ```bash cd /mnt/swarm-data/nextcloud/data rm -rf *-old-ssd ``` This is what actually reclaims the ~332GiB on `swarm-data`.
Author
Collaborator

Done — closing.

Verified on the filesystem: no *-old-ssd directories remain under /mnt/swarm-data/nextcloud/data, and swarm-data has drained to 1.7 GiB stored (from ~31 GiB), so the space was reclaimed.

Overtaken by events in the meantime: Nextcloud's data moved again on 2026-08-08, from hdd-bulk to the new swarm-data-ec EC 4+2 pool, and the USB HDD tier was decommissioned entirely (pool deleted, OSDs removed, drives pulled). Current state is data on swarm-data-ec — 408 GiB stored / 601 GiB raw, a 1.47x ratio — confirmed by reading ceph.file.layout.pool on a real file. That migration's own rollback copy (data-hdd-old) has also been deleted.

Done — closing. Verified on the filesystem: no `*-old-ssd` directories remain under `/mnt/swarm-data/nextcloud/data`, and `swarm-data` has drained to **1.7 GiB stored** (from ~31 GiB), so the space was reclaimed. Overtaken by events in the meantime: Nextcloud's `data` moved again on 2026-08-08, from `hdd-bulk` to the new `swarm-data-ec` EC 4+2 pool, and the USB HDD tier was decommissioned entirely (pool deleted, OSDs removed, drives pulled). Current state is `data` on `swarm-data-ec` — 408 GiB stored / 601 GiB raw, a 1.47x ratio — confirmed by reading `ceph.file.layout.pool` on a real file. That migration's own rollback copy (`data-hdd-old`) has also been deleted.
Sign in to join this conversation.
No milestone
No project
No assignees
1 participant
Notifications
Due date
The due date is invalid or out of range. Please use the format "yyyy-mm-dd".

No due date set.

Dependencies

No dependencies set.

Reference
copper/ccnet-prod-devops#37
No description provided.