Products > Computers
Strange hard disk performance issue
Halcyon:
I have a bunch of HGST 2TB 7K4000-series SATA drives (exact model is HUS724020ALA640) which were pulled from either a NetApp or IBM server some time ago (I don't remember which).
I've been attempting to securely erase them, but have found the performance is absolutely shocking, like abnormally so. Chopping it up to a potentially dead drive, I tested a few of the others, same thing. They seem to be reading/writing OK but just extraordinarily slowly (like it would take 3+ days just to format 2TB). Obviously something isn't right here.
My only hypothesis so far is that the drives have some kind of custom firmware from NetApp/IBM to prevent disk shucking and re-use in other machines and/or to ensure their customers only buy their "special" disks, but this is just pure speculation at this point.
This is not something I've ever run into before, particularly on SATA disks (these aren't even SAS). Has anyone got any thoughts? I have about 10 of these drives and would rather not shred them if they are still perfectly good.
MrMobodies:
Have you tried MHDD (dos) to see if the sector readings are indeed slow?
It reminds me a bit of Connor drives that I found very slow.
Back in 2005, I brought 4x 36GB 10,000 rpm scsi drives from an Ebay seller.
I connect them up, formatted all them and two failed.
I told the seller this and refunded me.
Now when I say formatted I mean by the controller before joining them on raid.
I noticed that the two that failed were HP branded on the sticker whereas the others were Seagate. Exactly the same drive, model number, shape and pcb board no difference in appearance.
JeremyC:
--- Quote from: Halcyon on January 25, 2023, 12:11:51 am ---I have a bunch of HGST 2TB 7K4000-series SATA drives (exact model is HUS724020ALA640) which were pulled from either a NetApp or IBM server some time ago (I don't remember which).
I've been attempting to securely erase them, but have found the performance is absolutely shocking, like abnormally so. Chopping it up to a potentially dead drive, I tested a few of the others, same thing. They seem to be reading/writing OK but just extraordinarily slowly (like it would take 3+ days just to format 2TB). Obviously something isn't right here.
My only hypothesis so far is that the drives have some kind of custom firmware from NetApp/IBM to prevent disk shucking and re-use in other machines and/or to ensure their customers only buy their "special" disks, but this is just pure speculation at this point.
This is not something I've ever run into before, particularly on SATA disks (these aren't even SAS). Has anyone got any thoughts? I have about 10 of these drives and would rather not shred them if they are still perfectly good.
--- End quote ---
I would try connect the drive to a Linux machine and try write zeros.
You must be root to do it.
Assuming that this drive is /dev/sdb:
dd if=/dev/zero of=/dev/sdb bs=10M
Detailed information about "dd": https://www.gnu.org/software/coreutils/manual/html_node/dd-invocation.html#dd-invocation
edpalmer42:
1. Use a program like CrystalDiskInfo to check the SMART parameters to see if the drives are just worn out. Look for 'uncorrectable sectors' or 'reallocated sectors'.
2. When you say 'format' do you mean a Windows format without the quick format option? If so, you could be bumping up against various OS limitations. Either use the previously mentioned dd under Linux or a utility program from one of the drive manufacturers. I always use Western Digital's Data Lifeguard program. They've discontinued it, but it's widely available at various download sites.
3. Do you have any security software running that might be interfering with the format process?
4. How are you connecting the drive to your system? Are your drives connecting at SATA 2 or 3 speeds? This drive has a sustained throughput of 171 MB/s which translates to ~1.4 Gbps. If your drive is connected through a USB2 or SATA1 connection your performance will be poor, but shouldn't be as poor as you describe. CrystalDiskInfo will tell you what type of connection you have.
Halcyon:
SMART checks out. In any case, I'm experiencing the same issue across multiple (identical) drives.
The issue occurs regardless whether I connect it via a USB interface or directly on the SATA controller. It also happens across multiple machines.
Windows format, or a utility like DBAN or MHDD produce the same slow results when erasing (about 8 MB/sec). Strangely the MHDD "scan" option appears to read the surface at proper speeds.
Doing some further investigation as we speak.
Navigation
[0] Message Index
[#] Next page
Go to full version