I purchased utilized business HDDs low-cost and found out the difficult method why they’re dangerous

Restaurants

Released Sep 11, 2026, 6:00 PM EDT

Umair Khurshid is an innovation author and designer with a strong concentrate on Linux, FreeBSD, cloud facilities, and automation. Before concentrating on composing, Umair worked as a designer and DevOps engineer structure and automating cloud-native systems.

Purchasing utilized business storage is among those things that makes ideal sense when you take a look at the rate and ends up being more complex once you really put the hardware into a maker. Business HDDs that initially cost a little fortune can appear on the utilized market for remarkably little cash. For a homelab, where storage capability is constantly helpful and budget plans are seldom limitless, it is challenging to neglect those rates.

<p>I just recently purchased a batch of 8 utilized 2 TB Seagate 3.5-inch SAS disk drives and one HGST Ultrastar DC SN200 after discovering a seller providing them inexpensively. The seller declared that the SAS drives were extra parts and had actually hardly been utilized apart from checking their performance. For a homelab, that seemed like precisely the type of offer worth gambling on.</p>    2 drives had a non-standard sector format

The drives had actually formerly been utilized with hardware RAID

2 of the drives appeared to have actually formerly become part of a hardware RAID system. They had actually been formatted with a non-standard sector size, utilizing 520-byte or 528-byte sectors instead of the 512-byte sectors that a standard HBA anticipates. This prevails with some business storage systems. The extra bytes can be utilized by the RAID controller for metadata or checksum info while the storage system provides a traditional rational sector size to the os.

That setup ended up being an issue when I linked the drives straight to my HBA. They had actually operated in the storage environment they were initially set up for, however they would not work properly in the direct-attached, software-managed setup I was constructing.

<p>The option was a low-level format that returned the drives to a basic sector size. I utilized <code>sg_format</code>which becomes part of the <a href="https://linux.die.net/man/8/sg3_utils" rel="noopener noreferrer nofollow" target="_blank"><code>sg3_utils</code></a>  bundle, to carry out the reformatting.</p>    


















                                                                                        <picture>
    </picture>

The procedure is much slower than a typical filesystem format. Depending upon the drive, the low-level format takes approximately 2.5 to 3 hours per TB. With big business disks, that can turn what at first appears like a basic compatibility issue into a whole afternoon of waiting. When the format finished, the drives might be utilized generally through the HBA.

The specific SN200 drives were a various experience

Low power-on time didn’t make them simple to utilize

                                                                                        <picture><source media="(max-width: 480px)" data-srcset="https://i0.wp.com/media.pmnsports.co.uk/2026/09/localimages/nvme-smart-report.jpg?q=49&fit=crop&w=500&dpr=2"><source media="(max-width: 767px)" data-srcset="https://i0.wp.com/media.pmnsports.co.uk/2026/09/localimages/nvme-smart-report.jpg?q=70&fit=crop&w=800&dpr=1">
    </source></source></picture>

I have actually constantly been more careful about purchasing pre-owned flash storage, and my experience with Western Digital’s Ultrastar SN200 series did little to alter that. The specific SN200 drive I purchased was a U. 2 NVMe gadget. The seller declared that the flash storage had actually been powered on for less than 12 days, which at first sounded exceptional, however I attempted utilizing it with Debian. It could not appropriately acknowledge the gadget, although smartctl might interact with it. The drive determined itself as a Cisco-rebranded gadget utilizing personalized firmware.

<p>Getting the drive working ended up being significantly more complex than the power-on figure recommended. I ultimately needed to handle firmware, vendor-specific tools, various os, and motorist concerns before getting the gadget into a functional state.</p>    <p>Even then, the issue wasn't entirely fixed. Throughout later screening, the drive might go back to Diagnostic Mode after numerous restarts, with the information on the gadget ending up being not available. That made it inappropriate for the caching work I had actually prepared.</p>    

The seller had actually noted the drive as non-refundable, however after I described the issue, they consented to accept the return. I would not take that experience as proof that SN200 drives are naturally bad. There was an issue with the specific drives I purchased, however it did demonstrate how little a power-on-hour figure can inform you about a previously owned business NVMe gadget.

The genuine expense is the screening

Low-cost drives need time before they make your trust

                                                                                        <picture><source media="(max-width: 480px)" data-srcset="https://i0.wp.com/media.pmnsports.co.uk/2026/09/localimages/seagate-smart-health-check-in-debian-terminal.jpg?q=49&fit=crop&w=500&dpr=2"><source media="(max-width: 1023px)" data-srcset="https://i0.wp.com/media.pmnsports.co.uk/2026/09/localimages/seagate-smart-health-check-in-debian-terminal.jpg?q=70&fit=crop&w=825&dpr=1">
    </source></source></picture>

The most significant lesson from purchasing pre-owned business storage is that the purchase does not end when the bundle shows up.

<p>Checking becomes part of the expense, for HDDs, I wish to examine SMART info, validate firmware, run full-disk composes and checks out, view temperature levels, and take a look at latency under continual work. If I'm purchasing a number of drives, I wish to evaluate them separately before putting them into a variety.</p>    

That matters a lot more with big ZFS swimming pools. A stopped working disk implies a replacement and a resilver, and a long resilver puts continual load on the staying drives. If those drives are likewise old, one failure can expose another weak disk at precisely the incorrect time.

ZFS offers checksums, redundancy, scrubs, and outstanding presence into storage issues, however none of those functions makes doubtful hardware dependable. Redundancy provides you space to recuperate from failures. It does not get rid of the possibility of those failures.

A backup stays essential no matter how costly or trustworthy the drives are. I still believe utilized business HDDs deserve purchasing

My experience didn’t persuade me to stop purchasing secondhand business HDDs. The batch I purchased turned out to be a quite excellent offer after comprehensive screening. The drives other than the 2 utilized in hardware RAID had low power-on hours, little taped compose activity, and no considerable irregularities throughout duplicated tests.

<p>The SN200 experience was much less enjoyable, however it likewise showed why I deal with utilized business NVMe in a different way from utilized HDDs. A drive can have extremely little taped usage and still present firmware, compatibility, or functional issues that aren't apparent from its health data. That is the genuine threat with previously owned business storage. You aren't merely purchasing a drive at a discount rate. You're purchasing hardware with an unidentified history and taking obligation for developing whether it appropriates for your system.</p>