All new Registrations are manually reviewed and approved, so a short delay after registration may occur before your account becomes active.
'Enterprise' HDDs : Share your power_on_hours stats and horror stories
HGST is probably in the "can't go wrong" category according to backblaze datadumps.
But since they're pricey and hard to get a hold of , what kind of powered_on_hours are you/providers here seeing from other brands of Enterprise HDD like:
Toshiba, Constellations, WD-RE3,REr, others
If members/providers could paste your model#/batch and powered_on_hours from your colo/rental servers from Hetzner, online, kimsufi , ovh, etc that would be useful for future decision-making via Google.
Also mention the hours at which you know a particular brand+model failed in the past for you. (especially silent deaths without proper SMART feedback )
Don't share stats of consumer models like seagate barracuda,momentus, WD blue, black, etc
Don't paste entire smartctl output, just the relevant model name+id and power_on_hours value
EDIT: and Reallocated_Sector_Ct value if non-zero.
else if you just have anecdotes about crazy long runtime hours with a particular model, go ahead and share. It could help people searching for it on Google
All entries in this thread should be taken as anecdotal and the sampling size is nowhere near that of Backblaze.
Comments
I'm using many drives that are still going strong with no errors and 50K+ poweron hours
Heymman Servers' HDD Seagate Constellation.2 (SATA) ST9250610NS
Power_On_Hours:
34739
34855
34739
34856
Four years in average. Are they too old?
I see big numbers on Seek_Error_Rate, Raw_Read_Error_Rate and Hardware_ECC_Recovered. Not sure what that means.
Reallocated_Sector_Ct is 0.
LT DEALS 1701.3 online.net e3 1240 €20
Device Model: ST1000NM0033-9ZM173
9 Power_On_Hours 0x0032 061 061 000 Old_age Always - 34625
Device Model: ST1000NM0033-9ZM173
9 Power_On_Hours 0x0032 061 061 000 Old_age Always - 34622
Hetzner Finland
Hetzner Helsinki
Edited with manufacturer and model
Hetzner Finland (Western Digital Re)
Western Digital Re, same as @beagle.
Sure here is some Western Digital Re at Hetzner Finland,
E3-1246v3:
i7-4770:
And here is HGST Ultrastar 7K6000 at OVH BHS,
E3-1270v6:
HDDs where brand new when I got the server few months ago.
Kimsufi KS-3C (i5-3570S):
HGST Ultrastar 7K4000
9 Power_On_Hours 0x0012 098 098 000 Old_age Always - 17691
No reallocated sectors, no horror stories, yet.
9 Power_On_Hours 0x0032 100 100 000 Old_age Always - 77050
5 Reallocated_Sector_Ct 0x0033 253 253 010 Pre-fail Always - 0
196 Reallocated_Event_Count 0x0032 253 253 000 Old_age Always - 0
Model Family: SAMSUNG SpinPoint P120
Device Model: SAMSUNG SP2504C
Hetzner Falkenstein still going strong
KS-3C BHS promo server (got it at 5000 hours or so?):
Online.net C2750 promo server (got it at 23000 hours or so?):
Both are doing just fine. Have had other dedicated servers with both providers before, old or new disks, never really had an issue, and hopefully never will.
Hetzner Finland
Online.net 20Eur E3v3
All drives had SMART extended test performed last week and Passed.
Auction server from Hetzner. I've had it for 1.5 years now, no problems whatsoever. The drives are Seagate Constellation ST33000650NS. Both drives reporting the same power-on hours and zero reallocated sectors.
sda:
9 Power_On_Hours 0x0032 054 054 000 Old_age Always - 41141
sdb:
9 Power_On_Hours 0x0032 054 054 000 Old_age Always - 41141
Dacentec
sda
sdb
Have 30710 with HGST and going strong
HDD poweron number never bothers me. I actually like higher numbers. With HDD, you can sort of predict when it's gonna fail.
Now, SSD scares me because they just go out with a 98 year-old grandma.
Colo
OVH E5 special
Hetzner Finland
Hitachi/HGST Ultrastar 3.5 years on both drives
From my 4x3TB Hetzner Server in their Nürnberg DC:
The other three HDDs are just Desktop HDDs with 25000-38000 Power_On_Hours
OVH GRA1 - HGST Ultrastar 7K4000 in Raid 10
Reallocated_Sector_Ct all 0.
... Company I worked for, and wasn't the server admin.
So we had our most important server, which contained most of our important information, running on RAID 5. This was done via a HP controller. Everything seemed fine over the years, and then one day we noticed a hard drive showing a red light. I reported this, but nothing was done for like 200+ days or something. At the time I didn't know we were running raid 5 on this server either. So finally the server admin got the hard drive, and the gumption to install the replacement hard drive.
... I think we all know what comes next, a second hard drive failure during the rebuild process. Instantly he panics, and starts an angry mutter. Don't worry guys we have backups, right? Well... while he waited for the rebuild to of course, fail and ran around panicking. Attempting to search for those "backups" on his custom solution.
In all thankfulness we did have those backups. Things we going to be good, and that would have been the close of our story. Assuming of course anyone tested those backups. Of course they weren't good, and had to restore to a backup almost 4 months out of date. So now my server admin is even more jaded. New hard drives were bought, and out went the long process of rebuilding the whole array, and data on it.
Did I mention he did this on the last day of a holiday, and people would be back the next day? Don't worry, it only took us like a month to get things back in order.
And now... you know why you have; replacement hard drives on hand for crucial systems, backups, tested backups, and time to properly rebuild an array.
We have 20+ ST33000650NS, 60+ ST4000NM0033, 20+ ST2000NM0011 in our dc. Most of reallocated sectors, if found on those drives, are logical bad sectors and were fixed. Only drives failed are 2xST4000NM0033, one at 23020 hrs, one at about 30000 hrs (dead and could not be detected).
Did somebody notice those >77000 power hours with 0 reallocated event counts of my Samsung Spinpoint.. ?
Not sure if Spinpoints fall under Enterprise.
I have a Spinpoint F1 and F3 too under my desktop's hood. (one is 7200rpm and 5400rpm. I forget which F'x' is which)
I'd call them prosumer drives at best. (better than the rubbish seagate was putting out in 2009-2012 )
They're haven't had a single re-allocated sector yet (only 2yrs power_on_hours though)
Some of those "rubbish" Seagate drives are actually pretty good. I have one from around that era with 53,000 hours on it and it is still going strong.
are they too old? how many power on hour i should consider to change hdd.
Here's a list of all the enterprise drives from boxes that I rent including a couple running in my basement. I'll preface this by saying, I've yet to have a drive fail on me yet, so can't advise as to which are the most reliable, etc. No drives below have any reallocated sectors.
Choopa Custom Game Server (NYC) -
SoYouStart GAME-2 (BHS) -
DatabaseByDesign E3 Special (Philadelphia, PA) -
DatabaseByDesign Cheap Fucks Special (Philadelphia, PA) -
VirMach Top Provider Poll Prize (Buffalo, NY) -
Heymman Servers E3 WHT Special (Chicago, IL) -
Basement Server #1 -
Basement Server #2 -
Hetzner SB27 i7-4770 2x2TB Ent. HD (Helsinki DC)
Western Digital WD2000FYYZ 2TB
20670 hours
30214 hours
Hetzner SB27 i7-4770 2x2TB Ent. HD, in Helsinki like @roshan91.
Hetzner e3 1246v3 helsinki
Model Family: Western Digital RE4 (SATA 6Gb/s)
Device Model: WDC WD2000FYYZ-01UL1B3
9 Power_On_Hours 0x0032 086 086 000 Old_age Always
- 10626
9 Power_On_Hours 0x0032 098 098 000 Old_age Always
- 1883
I think i got lucky there