New on LowEndTalk? Please Register and read our Community Rules.
All new Registrations are manually reviewed and approved, so a short delay after registration may occur before your account becomes active.
All new Registrations are manually reviewed and approved, so a short delay after registration may occur before your account becomes active.
Old disk in server
Hi,
Recently I got a disk failure, got it replaced by DC guys and re-synced , howeer, from smartctl , I can see that disk has Power ON hours of about 48374 (roughly about 5.5 yrs power on ).
But rest of SMART RAW stats are good, like Reallocated_Sector_Ct is 0, Current_Pending_Sector 0, Offline_Uncorrectable 0 . The disk show only 21 Power_Cycle_Count.
Smartctl short and long tests passed. But I am still concerned, do you think device being very old, should I ask to get it replaced by DC guys or old drives are fine to go if SMART stats are good ?
Comments
Depends on what you pay.. Paying for lowend = getting lowend.
But normally if they replace a broken disk, they shouldn't replace it with old one..
48374 hours = not good at all for an HDD. This is just waiting for a new failure.
Depends on the drive, and as said, how much you pay.
I've had drives replaced in the past with used drives, although not with that many hours.
I have a dedicated box still going fine with about 65k hours on both drives with a power cycle count of about 12.
You could ask the provider to change it for peace of mind, but they may not if the drive is otherwise healthy.
I have a dedi from hetzner that has two drives in raid with 50k hours. They seem healthy but you get what you pay. I am paying 21.8 eur and that is pretty much low price. If you are paying more then consider switching to company that will rent you brand new servers.
In general I would say as long as the item you rented is working fine, you have nothing to complain about. Of course if it dies within your rental period, it needs to be replaced for you.
When you rent a car you may get a nearly brand new one but also one with 100k+ kilometers. The car rental company needs to ensure it does not break down (or replace it if it does) but in general you do rent "a car" - not "a new car".
Same here. You rented a server with some hard disks, probably without any specifications about the age of the drives.
Always keep monitoring your disks health so you can raise a ticket as soon as some vital SMART features start increasing.
Yep, I understand why you may not like it, but if it works it works, if you dont like it...colo
My Hetzner auction server when I got it had 1 new Toshiba drive and 1 very old Seagate drive of a notoriously unreliable model, with some non-catastrophic SMART errors. I suspected it previously had two unreliable Seagates and one of them failed, so they replaced it and put the server into the auction. I opened a ticket but they said they don't replace drives just because they are old. They also said that if they did replace the drive, it might well be with another old one. Then the SMART error count in the old drive started increasing so I opened another ticket and that time they replaced the drive, with another new Toshiba. Both Toshibas are still working fine a few years later.
Which model may I ask? One of my auction servers has 2 x Seagate Constellation ES.2 in. They both have almost identical hours on but power cycle count is off by about 3 between them.
One did start increasing the error count, so I was preparing to ask for it to be changed but has since dropped back to 0 and has been fine since lol.
IF it is an expensive server, then, it should be provided to you with new or almost new disks. On a cheap server, the drives are never new. In fact, in most of cases, there are rather old. It is a compromise you have to do for the price.
Now, if SMART values are fine, then, don't worry about the drives. I have a couple of drives in a Dacentec server that have operated for more than 65000 hours and are still fine. They are in a RAID 1 environment and SMART values are still excellent with no bad sectors, no problems on spinning etc.
Besides, a HDD that is operating for 50.000 hours without any issues, has proven that is issue-free. New disks can fail or face issues in the first months or year of operating.
Monitor your disks, keep backups for your data and have a RAID 1 setup when possible for an extra layer of redundancy. And when you see errors in the SMART logs, then, ask your provider for a disk replacement.
Which unreliable Seagate disk? ST3000DM001. It was notorious and there was later a lawsuit about how awful it was. Lol, there is a wikipedia article, mostly about its unreliability:
https://en.wikipedia.org/wiki/ST3000DM001
It is a Barracuda model. Constellation is an enterprise series that I think was not affected by the problem.
Great news, thanks.