I buy spinning rust as bulk storage because they used to be reliable medium term storage devices. Now my top of the line Toshiba drive is down to like 80% helium capacity after 18 months. Who decided this was a good system and were there any adults in the room to explain how stupid this was?
>>106462049well maybe that was the point? planned obsolescence
>>106462049Don't you think the engineers were smart enough to account for any potential leaking of Helium through the seals or whatever they use, and OVERFILL / add some margins to the drive with helium? Even if some helium is leaking or gets soaked into the seal materials, there's still enough helium to keep the drive working inside.
I forgot why they even started filling them will heliumspinning resistance or inert gas for HAMR?
Why don't we just make a vacuum inside instead of any gas?
>>106462049idk some anon in another thread posted his helium drives are still at 100% after years
>>106462149just like helium makes your speech higher pitch, it also makes the discs higher rpm
You will own nothing
>>106462207he will own the helium free drive
>>106462113Maybe it’s just mine. The SMART stats are fucked too. Over a hundred reallocations. Did a surface scan with disk genius and zero sectors lower quality than “good”. Did a low level format and zero bad sectors. Just bad luck on my part I guess.
>remove helium filled hdd from pc>it floats away
>>106462167I think the gas helps create the gap.
>>106462273Oh right, otherwise the heads would just lay on the platters. I'm dumb.
>>106462285Ye olde Bernoulli effect
>>106462049Helium HDDs have been deployed enmass since like 201699% of cases of helium leaking is physical damage.Even the super early helium HGST 8TB HDDs didn't have problems with leaking but outgassing internally from being ran hot as fuck at over 55C+Plenty of HDDs have failed but helium was never a main cause. It can be a failure mode through no fault but the manufacturer but it hasn't been a thing that mass wipes out an entire generation of HDDs.It should also be noted that for all the problems with any particular generation of HDDs, the hyperscaler I get this info from has never decided to outright stop using them.
>>106462049>like 80% helium capacity after 18 months.You treat your hardware like shit, don't you?
>>106462437No. It is in a well cooled case which has never been moved while running. And I’ve never put more than 10TB on this 20TB drive. It’s the only piece of hardware I own that is fucked up. Also fuck your own face.
>>106462488>Also fuck your own facehi jeetlon
>>106462237>Did a low level formatNo you didn't, because modern drives don't support that. Writing all zeroes is not even a format, stop calling it "low level format". Low level formatting was able to change the physical format and positions of sectors, that hasn't been the case since before PMR, sector size and location is encoded on the platter at the factory and all the drive can do is read it for tracking (and accidentally damage it leading to IDNF errors, but that's a different can of worms).
>>106462317One issue with helium drives is because of the helium they are effectively non-repairable. But on the other hand, if they fail in warranty you're probably not getting a refurb as replacement.
>>106462851Under no circumstance should "repairing" or recovering an HDD even be necessary.You're looking at spending $500-1000 to save a few hundred in a backup.
I have a drive with 8 reallocated sectors (one 4K block) but absolutely no failure indicators. No uncorrectable, no pending, no high flying, attributes all look normal, no error log entries at all. Has anyone ever seen this happen before? It's like the drive just decided to reallocate one block on a whim.
My drive twisted itself into a giraffe.
>>106462049My 11yo HGST is still at 100%
>>106462933It can happen.People usually interpret any bad sectors as being indicative of failure.In some cases it's true, bad sectors could be a symptom of something bigger like mechanical failure when they show up in conjunction with the drive also experiencing issues with reading and writing.If it's just a handful of sectors and the drive works fine it's likely just a surface defect.I have a SAS drive with some potentially reallocated sectors. The grown defects list started increasing but stopped at 147. ZFS has no complaints about the drive so it's probably something on the platter surface.
>>106462149>I forgot why they even started filling them will heliumit makes data centres less heavy
>>106462813Yes I know. It’s a “low level format”. Real deal low levels went out with RLL. Modern equivalents scan for errors, and mark the sectors bad. It’s functionally equivalent
>>106462933Yes, that’s exactly what I’m experiencing.
