HDD issues.

[Spork Schivago]

Re: HDD issues.

I have not used the Red drives, but I have some Re's, and I like them.

WD Red drives have a 3 year warranty. They are not really enterprise drives. WD says, "WD Red NAS hard drives are recommended for use in home and small office 1-8 bay NAS systems." WD rates them at 1 million hours MTBF. Here is one review of the 4tb Red drive.

WD Red Pro drives have a 5 year warranty.

WD Re drives have a 5 year warranty. They are enterprise drives. WD says they are rated for 2 millions hours MTBF and "Designed to handle workloads up to 550 TB per year".

Note that the BackBlaze reports do not indicate that enterprise drives have fewer failures in their use case (mostly being written to until they are full then spinning with few accesses until somebody needs to restore a backup). This does not mean they don't last longer for their intended purpose (24/7 data transfers).

Thank you for the correction! I've been out of the loop for a bit, working on other things besides PCs. Trying to do something different with my life and all that jazz.

[Spork Schivago]

Re: HDD issues.

The firmware doesn't "scatter" any data. The OS determines where the data are written. AV drives incorporate the ATA Streaming feature set which allows the drive to skip difficult sectors whereas a typical desktop drive will become bogged down with error recovery. The philosophy is that AV applications can tolerate missed video frames or audio dropouts.

Yeah, isn't it the filesystem that determines where the data gets written? Aren't there fragment free filesystems out there? There's so many of them it's hard to keep track. I kinda liked where Btrfs was going. I wonder if that's stable now. I think some Linux distro's ship with it by default.

[Spork Schivago]

Re: HDD issues.

The ROM does not decide where the data goes. No drive "scatters" data. The OS tells the drive to write to LBA n, and the drive then writes to LBA n. However, the drive has a "translator" which determines how the logical blocks are mapped to the drive's physical CHS addresses. Even so, the data are still laid down in serpentine fashion, in LBA order, except for the grown defects. Please read the Tracks and Zones article. It was written by a Seagate employee who is a highly regarded member of the data recovery community.

Oh man! I got my picture taken with the old VP of Seagate!! I put my arms around him like we was buddies and shit! I went to his daughter's wedding! She was marrying my friend and he picked me as his best man. They flew me (and I flew my now-wife) out to Australia for the wedding. It was insane. Never lived so rich!!! Just the motel alone we stayed in must of cost a grand a night! We went on this yacht for dinner, and they had a live band playing for us, it was insane! We ate out every night and the dad payed for it all! We found common ground, my home town, Corning. He said Seagate was trying to make a hard drive that had glass platters. They contacted Corning to make the glass but there was something wrong with the design and they couldn't work it out. I can't remember what. Maybe they where too heavy or they shattered at the fast speeds? I dunno. He was super intelligent though. I'll try to find the picture and if I can, I'll post it here.

[fzabkar]

Re: HDD issues.

I'm not aware of any modern file system that doesn't allow for fragmentation, but then I haven't looked much beyond Microsoft. In fact that last time I saw a file system that would only write files in contiguous space was in the 1980s.

PATA has a speed limit related to the selected UDMA mode whereas SATA can transfer data at a much faster rate, especially for SSDs. You would only use IDE compatibility mode for legacy OS-es or in the absence of a suitable SATA driver. For example, some HDD manufacturers' firmware updates are delivered via DOS based utilities which sometimes have trouble communicating in SATA mode.

As for glass platters, the data recovery guys often relate horror stories about shattered platters, or platters that have suffered a prolonged head crash, to the point that all the oxide has been scraped off and the platters have become transparent.

https://upload.wikimedia.org/wikiped...iled_Disks.png
http://www.nitroware.net/images/stor.../deskstar2.jpg

[Spork Schivago]

Re: HDD issues.

I'm not aware of any modern file system that doesn't allow for fragmentation, but then I haven't looked much beyond Microsoft. In fact that last time I saw a file system that would only write files in contiguous space was in the 1980s.

PATA has a speed limit related to the selected UDMA mode whereas SATA can transfer data at a much faster rate, especially for SSDs. You would only use IDE compatibility mode for legacy OS-es or in the absence of a suitable SATA driver. For example, some HDD manufacturers' firmware updates are delivered via DOS based utilities which sometimes have trouble communicating in SATA mode.

As for glass platters, the data recovery guys often relate horror stories about shattered platters, or platters that have suffered a prolonged head crash, to the point that all the oxide has been scraped off and the platters have become transparent.

https://upload.wikimedia.org/wikiped...iled_Disks.png
http://www.nitroware.net/images/stor.../deskstar2.jpg

I guess the way I worded it was incorrect. Technically, fragmentation in Linux can happen, but the file system, like ext 2, 3, 4, etc, will automatically take care of it when it happens. Here's a decent article that talks about why we don't need to defrag in Linux and why Linux distros don't come with defragmentation

http://www.howtogeek.com/115229/htg-...defragmenting/

It's a little old, 2012 I think, but it's still valid. I'm not sure how the newer file systems in Linux deal with fragmentation, like Btrfs (B-Tree File System or Binary Tree File System).

[Per Hansson]

Re: HDD issues.

I split off the SSD discussion into this new thread:
https://www.badcaps.net/forum/showthread.php?t=49114

[Spork Schivago]

Re: HDD issues.

Thank you.