For me it's power issues. I've had two Raspi4s die on me (ssd/sd card corrupted) and I think each time it was due to an ill-timed power outage or momentary power dropout.
So Raspi4 + UPS is probably maybe okay. Just underpowered and overpriced.
Mini PC + UPS def robust and reliable.
Old laptop w/built-in battery just about ideal for people that just want to keep things simple.
For data backups, my strategy is moving toward:
* virtualizing containers via ProxMox
* dual mini-PC ProxMox cluster
* nvme system drive in each mini PC
* 2 TB ssd data drive for each
* 8 TB mirrored NAS spinning hdds storing scheduled backups from the 2TB data drive AND the ProxMox container image/snapshots.
So any system nvme can fail, any 2TB data ssd can fail, and either of the physical mini PCs can die and I'd still be able to rebuild everything as long as one of the mirrored hdds survives.
And all of this is kind of just tech ops LARPing as I really don't need this level of redundancy and failover. I don't run LN so the vast majority of my node data is totally unimportant (lost my node data? Meh, IBD again).
KeithMukai / Keith Mukai
npub1tv…77wn2
2023-09-23 15:59:36
in reply to nevent1q…e7ck
Author Public Key
npub1tv8gmfhalwnxxquxjzeh6gtdsdz6vg7vx0s3rt7s7uuw6aujh32qn77wn2Published at
2023-09-23 15:59:36Event JSON
{
"id": "3dcc94f1bb4dbdad36dc1bc6a30f7107890ca8a5a59b53cec0414eac180e7d53",
"pubkey": "5b0e8da6fdfba663038690b37d216d8345a623cc33e111afd0f738ed7792bc54",
"created_at": 1695484776,
"kind": 1,
"tags": [
[
"e",
"53e3839077f6e272ef16c14b6bd1f30ca6996aecd43c1ae41d28066a47a5c2d3",
"",
"reply"
],
[
"p",
"738f69184aeda675002b687fe47c8e9e2f7b1a267d6f9145b1193312f97c18ef"
]
],
"content": "For me it's power issues. I've had two Raspi4s die on me (ssd/sd card corrupted) and I think each time it was due to an ill-timed power outage or momentary power dropout.\n\nSo Raspi4 + UPS is probably maybe okay. Just underpowered and overpriced.\n\nMini PC + UPS def robust and reliable.\n\nOld laptop w/built-in battery just about ideal for people that just want to keep things simple.\n\nFor data backups, my strategy is moving toward:\n* virtualizing containers via ProxMox\n* dual mini-PC ProxMox cluster\n* nvme system drive in each mini PC\n* 2 TB ssd data drive for each\n* 8 TB mirrored NAS spinning hdds storing scheduled backups from the 2TB data drive AND the ProxMox container image/snapshots.\n\nSo any system nvme can fail, any 2TB data ssd can fail, and either of the physical mini PCs can die and I'd still be able to rebuild everything as long as one of the mirrored hdds survives.\n\nAnd all of this is kind of just tech ops LARPing as I really don't need this level of redundancy and failover. I don't run LN so the vast majority of my node data is totally unimportant (lost my node data? Meh, IBD again).",
"sig": "5aef2a3a6480160b7fb67967bd6db3aad2b3b370ad0d2d79227fcc341df2131604b2cc69eb9fa320a1ee230f8079061f7884bdace802cbd732848258c93e3442"
}