- Fractal Meshify 2 XL
- 10x 12TB drives in RAID6/Z2
- AMD 5950X (32 threads, 20 4.0X lanes)
- 64GB RAM
- LSI 9300-16i in HBA
- GTX3090 for LLM (second 3090 needs new mobo due to lanes)
- Fractal Meshify 2 XL
- 10x 12TB drives in RAID6/Z2
- AMD 5950X (32 threads, 20 4.0X lanes)
- 64GB RAM
- LSI 9300-16i in HBA
- GTX3090 for LLM (second 3090 needs new mobo due to lanes)
Lasset die Daten kommen. 😬
Lasset die Daten kommen. 😬
Die fehlt in einem RAID6 und es wäre echt cool, eine neue zu kriegen... 😅
Deren Support: unnötig, schickt nur Textbausteine. 😠
Die fehlt in einem RAID6 und es wäre echt cool, eine neue zu kriegen... 😅
Deren Support: unnötig, schickt nur Textbausteine. 😠
But it turns out that triggers a per-boot benchmark(!) of RAID6 algorithms. On every boot.
But it turns out that triggers a per-boot benchmark(!) of RAID6 algorithms. On every boot.
raid6+btrfs holds subvolume with postgres database.
On boot, systemd creates a zram block device and initializes it with btrfs. Then send()s the latest snapshot of the postgres subvolume to the in-ram btrfs.
raid6+btrfs holds subvolume with postgres database.
On boot, systemd creates a zram block device and initializes it with btrfs. Then send()s the latest snapshot of the postgres subvolume to the in-ram btrfs.
I could upgrade a couple of drives for a lot more capacity (4, 6, 8, 8, 8 in it atm, Raid6 treating all as 4TB), or go a new NAS with all new drives.
What would you do?
I could upgrade a couple of drives for a lot more capacity (4, 6, 8, 8, 8 in it atm, Raid6 treating all as 4TB), or go a new NAS with all new drives.
What would you do?
Every 5 minutes, the subvolume in RAM is snapshotted and send() to the btrfs on the raid6
snapshot retention on your typical hourly/daily/weekly/monthly rotation.
I get disgustingly fast postgres with rolling backups and a 5m data loss window.
Every 5 minutes, the subvolume in RAM is snapshotted and send() to the btrfs on the raid6
snapshot retention on your typical hourly/daily/weekly/monthly rotation.
I get disgustingly fast postgres with rolling backups and a 5m data loss window.
🔸 Now adds RAID6 & JBOD to existing RAID0/1/5 support
🔸 Smarter disk combining and future-proof expansion
🔸 Clean, intuitive UI for effortless setup
#ZimaOS #RAID6 #JBOD #NAS #SelfHosting #ZimaBoard #HomeLab #DataStorage #StorageSolutions #UPDATE #Linux
🔸 Now adds RAID6 & JBOD to existing RAID0/1/5 support
🔸 Smarter disk combining and future-proof expansion
🔸 Clean, intuitive UI for effortless setup
#ZimaOS #RAID6 #JBOD #NAS #SelfHosting #ZimaBoard #HomeLab #DataStorage #StorageSolutions #UPDATE #Linux
what are ppl using these days for cheap secondary archiving w/ infrequent access? S3 Glacier? some university thing? #psychscisky #neuroskyence 🧪
we have 20-30TB I'd really like to not lose in the event of a catastrophic failure
we've had an enterprise grade server with a 12-drive RAID array running 24/7/365 for the last 8 years. this week one of the 12 drives failed
I guess a hard drive can only spin so long. RIP
what are ppl using these days for cheap secondary archiving w/ infrequent access? S3 Glacier? some university thing? #psychscisky #neuroskyence 🧪
we have 20-30TB I'd really like to not lose in the event of a catastrophic failure