Avatar
iru@localhot $_
9eeca894b2e8830e05ec86f4d119de411a2dd3abcb65525db69c486f4afe5882
Technological optimism in current dark age.

Poland App Store return never? Might as well do web like bittr if Apple is being a jerk about it.

Hear me out. Block Clock’s Moscow time is actually a divergence meter. When it reaches zero it means we returned to the original world timeline.

I got this working with nostr:npub1u8lnhlw5usp3t9vmpz60ejpyt649z33hu82wc2hpv6m5xdqmuxhs46turz's lnurl-server.

nostr:note1uyalkausww8pfq55xtkvv6lp93q0aherjgnfeet4qldxehrs07ysnxyddd

Yes I have a proper node. I don’t want to expose it though since its running in my closet. I plan to expose the REST port to some Hetzner server running the web part and a simple invoice macaroon authentication.

Simple #lnurlp software? I don’t want to run a full fledged BTCPayserver or LNBits. Want something like lndhub-go.

Cool. What’s powering those models?

As for me I roll offline.

nostr:note1x5gtvskx0gmq6xheqlt58p28jksgft3g7f5zj9qatnue0c0ppksq0t76ax

Had no idea about unleashed.chat. Sadly Mixtral 8x7b is cucked by French “alignment” EU centric bullshit.

CAN CONFIRM I FORGOT MY CALCULATOR ON AND NVK SNEAKILY INSTALLED FREEBSD ON IT.

Yup small setup for a single person. Cheap and silent is better.

Not arguing with reliability. I want to say I literally never ever had a disk problem on my hand that affected me more than the annoyance of replacing or redeploying.

I keep getting nice work offers on the UK. Don’t really want to leave beautiful and based Polska for that crazy island lol.

brands have a specific set of manufacturing facilities and plants, in general, ok sometimes that overlaps but not often with long established brands do they use competitors, usually due to contractual obligations

a backup is worthless if it is stored on a device with poor quality control as the worse the QA standard, the more likely that when you come to rely on it, it's not working

fortunately, for the most part flash storage doesn't become unreadable when it fails, but spinning disks can gum up due to bad lubricants that have aged and the disk doesn't spin, or the media craps out leading to a higher error rate than the drive's FEC setup (which is not reed solomon like it is on ISO formated storage btw, also i think tape backups use RS FEC)

usually there is a cost premium associated with the level of QA that a manufacturer requires of their devices, and this QA level usually entails stress testing a certain percentage of each production batch to establish a benchmark for the batch lifespan estimate and thus insuring the warranties are not likely to void very often until after the devices fail

this is why it's worth spending a little extra on a better quality device with a lower return rate because it likely means they sacrifice a larger percentage of production to quality assurance and that warranty period has more relation to reality than their budget for covering this cost

long warranty periods don't necessarily mean that the devices are likely to last longer, just that the sales volume enables them to cover a higher failure rate

i don't know what WD's QA benchmarks are these days but me and semisol both avoid this brand because in the days of spinning disks, they were notorious for failing before the warranty was out, or not long after it, whereas other devices, sometimes even with shorter warranties, would last far longer in practise.

this is a practise known as "loss leading" in the industry and tends to be used by manufacturers that are aiming at cheapskate users like you

you can risk your data with it but it's very likely that you pay for your decision to go with a cheapskate company with a 3 failures at once situation and you lose data completely

Do not stress on hardware when software and configuration is what matters. Data won’t be corrupted because a modern file system does checksuming and self healing of the blocks. SMART monitoring will alert you when things are about to go bad. There won’t be a loss of data considering backups are off site. TBH actually important part of this setup that needs to be taken care of are 371MB of mostly iddle data.

The nuance take you guys don’t get is that most people aren’t running a service datacenter where every transaction matters. It’s not being a cheap stake. It’s about being smart. Not relying on parts that are harder and more expensive to come by when you can just scale horizontally are a good thing.

It broke? Who cares? Restore backup, resync chain if needed and have it running on a few hours while you are hunting hardware, potentially on eBay hunting for used discarded corporate hardware. Vendor will RMA right? Save money and get some extra sats.

> Backup is worthless if its stored on a device with poor quality control

Why would you store backups on the same device?

> avoid this brand because in the days of spinning disks

SSDs are are a totally different beast. In this case the NAND chips are built by Sandisk and controllers are Marvell. If it breaks I will get another and get stuff running for no price.

You also don’t need to rely on anecdotal experiences from 20 years ago. There is Blackblaze that publishes reports every year.

But should I care? It has a 5 year warranty and redundancy. By that time a larger disk will be required to store the blocks.

People should focus on backups rather than brands.

Replying to Avatar Jameson Lopp

Mushrooms can be cheap depending on where you live.

I’m not running a service provider. I’d rather scale horizontally plus backups and have the replacement at my door the next day. RAID redundant array of inexpensive disks after all.