I once saw something like that when the disk had a leftover GPT partition and grub got confused and refused to boot, and fdisk A refused to partition because it complained about GPT, and gdisk claimed that the disk had MBR partition scheme. dd-ing zeroes over the first couple MB (8, IIRC, and overkill) got rid of the partition, made grub happy and allowed fdisk to partition the disk without complaining about GPT. On Oct 9, 2017 17:24, "D. Hugh Redelmeier via talk" <talk@gtalug.org> wrote:
| From: James Knott via talk <talk@gtalug.org> | | On 10/07/2017 08:53 PM, Howard Gibson via talk wrote:
| > Best Buy offered me a newer Seagate at a slightly lower price but one | > is claimed explicity to support Linux, and it supports some older | > protocols. When I told the people at the store I wanted MBR, not GTP, | > they just stared at me. | | That's proof store clerks are not allowed to know what they're talking | about. What does a disk drive care about what OS is on it? All it does | is read/write data from specified locations on the disk.
The PC world is full of hardware and software of various levels of conformance to various "standards".
As a consequence, when someone builds a new component, they often try to trick old components into working. These tricks can sometimes be indistinguishable from bugs.
Here are some such HDD tricks that I remember:
- jumpers to have drives report smaller capacity to overcome OS or BIOS limitations
- ATA drives that will pretend to have any geometry (picked from some clues during booting, I think). Of course they actually had no geometry
- "hybrid" drives (HDD with modest SSD cache) that "knew" OS disk access patterns and allocated SSD to match
- drives with 4k sectors that faked 512 sectors (to various extents) to allow old BIOSes to boot from them
- drives that claimed their on-disk cache was flushed before it was.
But I don't remember any disks that were not supported by Linux. --- Talk Mailing List talk@gtalug.org https://gtalug.org/mailman/listinfo/talk