Understanding SMART for the not so smart

  • Please forgive me if this has been posted a thousand times before. I've searched through the Forum and found a lot of posts on this however I still cannot decipher what actions if any I can take to resolve or smart/drive is doing its job or if I should get a new Drive.I decided to post my info and have the professionals look at it any insight would be helpful and appreciated.


    I purchased this drive off Amazon everything seemed to work good, so I bought another one so I have two of the exact same model however the first one started showing bad sectors. I know this is topic that is discussed a lot and I'm sorry for another paranoid person worrying about its Drive post, but people get paranoid about their drives and I am one of them :rolleyes: . I set up smart in the GUI and set up some schedules for once a month (btw is that okay or what is recommended for smart scheduling) The second hard drive I ordered is working fine and so is the SSD that runs OMV.
    Once I got my first email from smart saying I had 24 Currently unreadable (pending) sectors It got me concerned I did do a lot of reading and sort of understand if I'm right it's waiting to read those sectors and remove the data that can replace and set the bad sector aside. I could be very wrong.

    This is my daily email.

    I received this one once after a reboot.

    Could someone explain to me Reallocated_Sector_Ct?

    And remaining and lifetime?
    smart.txt
    That is my info if i do smartctl -x


    Any insight would be greatly appreciated.
    Bruce

    My homebrew Nas
    OMV 4
    ASRock760GM-HDV
    AMD Phenom II X6 1065T
    Viper 3 16GB DDR3 1600MHz
    PNY 120GB SSD
    Two: Hitachi 4TB
    In a HP Pavilion case :D

    • Offizieller Beitrag

    Sorry I missed this thread. While there is no perfect indicator of a hard drive failure, a BackBlaze study showed the following SMART attributes to be prime indicators of a future drive failure.


    SMART 5 – Reallocated_Sector_Count.
    SMART 187 – Reported_Uncorrectable_Errors.
    SMART 188 – Command_Timeout.
    SMART 197 – Current_Pending_Sector_Count.
    SMART 198 – Offline_Uncorrectable.


    One or two in the raw counts, of the above, might be a matter of concern - something to watch. If they're incrementing upward, it's just a matter of time. The Extended off line test also supports that the disk is beginning to fail. It could be weeks or even months, but it could be tomorrow. No time should be wasted in getting data off the disk.

  • Sorry I missed this thread. While there is no perfect indicator of a hard drive failure, a BackBlaze study showed the following SMART attributes to be prime indicators of a future drive failu

    That's quite alright there's a lot of activity is on this forum. It has not risen since the initial notification I just know that I have a few bad sectors and honestly, when you get that notification the paranoia sets in. I get an email every day when it starts to change to a higher count of unreadable sectors then I'll start worrying thank you for your time. I do have a external backup of everything to be safe. I initially wanted to do two hard drives as raid 1 to start out with and an extra drive for backup. But one drive came in earlier than the other and I got a little antsy to set it up :) Here shortly I'm going to purchase some more drives and I'm thinking about doing raid 5.


    I used the OMV to share media throughout the house with Plex and for backing up and and some docker stuff. I'm going to automate my backups rsync In the future.
    I do have a couple of friends that have access to my Plex however, I think the most connections were to the server was three watching plex and there was no lag or hiccups during that time. Since then I have upgraded the CPU to 6 core processor. For the most part I personally watch my media through network sharing on a Vero 4k+.



    If I could pick your brain for a moment what would you recommend for raid? Now that you know basically what I use OMV for. My current setup is AMD Phantom II 2 x6 1065T 2.9 GHz, 8 gigs standard DDR3 1600MHz (PC3-12800) 120 gig solid state drive for the operating system in the board is a ASRock 760GM-HDV.


    Thank you for replying


    Bruce

    My homebrew Nas
    OMV 4
    ASRock760GM-HDV
    AMD Phenom II X6 1065T
    Viper 3 16GB DDR3 1600MHz
    PNY 120GB SSD
    Two: Hitachi 4TB
    In a HP Pavilion case :D

  • Just a quick update it's been a couple of months since smart notify me of the bad sectors and they have not risen. Everything seems to be working just fine I just take it as a couple of bad sectors.

    My homebrew Nas
    OMV 4
    ASRock760GM-HDV
    AMD Phenom II X6 1065T
    Viper 3 16GB DDR3 1600MHz
    PNY 120GB SSD
    Two: Hitachi 4TB
    In a HP Pavilion case :D

    • Offizieller Beitrag

    If I could pick your brain for a moment what would you recommend for raid? Now that you know basically what I use OMV for.

    First, if you have backup (at least for personal files that can't be replaced), you're way ahead of most. The only way to insure that you'll keep your data is to back it up, as in a second copy of everything you want to keep. Otherwise, it's just a matter of time....


    "RAID", for most home servers, doesn't make a lot of sense. While RAID solved a few problems back in the day, with today's enormous drives and alternate drive pooling methods, the value it once had has diminished. As examples, UnionFS or even symlinks are ways to pool drives that don't involve RAID.


    To look at one potential home use for RAID; I'm using ZFS in the equivalent of RAID1 for bit-rot protection, but it comes at a steep price of losing 1/2 of available disk space, characteristic of RAID1. Note that I'm not using the ZFS equivalent of RAID1 for drive "backup" or even for "availability". (RAID is not backup.) The zmirror is solely for bit-rot protection.
    ____________________________________


    If you want to pool drives for large file (video) storage, you might consider UnionFS. If you don't want to get into the details of UnionFS storage policies, the best policy for a large Video library might be "Most Free Space" which is not the default. If using UnionFS with 2 to 3 drives, along with the "Most Free Space Policy", I'd also add a SNAPRAID parity drive.


    There are lots of benefits with SNAPRAID; file, folder and drive restoration (to the state they were in as of the last SYNC) and bit-rot protection. (BTW: Routine Tasks like snapraid sync and snapraid -e fix and similar commands can be run manually from the GUI, but I've automated them in Scheduled Jobs. (I'm using UnionFS+SNAPRAID on a backup server.)


    The combination of the two UnionFS+SNAPRAID are better than RAID5 IMO. Since you have an SSD for the OS, I'd keep Dockers or VM's (if any) on the boot drive, not on the UnionFS mount point.

  • That makes perfect sense. Perhaps I won't worry about doing a raid. I do an offline backup manually of everything that is on the NAS to be extra safe well, I do that for everything all the pc in the house. I was thinking of doing raid just for the geek factor but the way you put it is unpractical and a waste money. You've given me a lot to read about time to do some research. UnionFS+SNAPRAID Looks very interesting Thank you for the reply I appreciate it.

    My homebrew Nas
    OMV 4
    ASRock760GM-HDV
    AMD Phenom II X6 1065T
    Viper 3 16GB DDR3 1600MHz
    PNY 120GB SSD
    Two: Hitachi 4TB
    In a HP Pavilion case :D

  • Thank you! I have used rsync before to do automatic backup to an external drive and was planning on using rsync to do my backups on the NAS. I believe you probably know the YouTuber Techno Dad Life he did a pretty good tutorial on it also. Eventually, when I have time I would like to set up all my computers to do rsync remotely to the server. One of my main things to right now is finding a decent case it looks like a Rat's Nest in the case it is an old HP Pavilion case lol and the space is getting limited :-).


    I'm planning on buying 1 Moore Drive 8 or 10 terabyte drive to store backups of everything and I'm go to up the ram to 32GB. Oh and I'm going to add a TV Tuner Card in the future.


    What I use the NAS for is for every person in the house has their own folder set up as their own personal backup and a general share folder for everyone to access just a general Windows networking I'm the only one that exclusively uses Linux as an everyday computer. I use Plex and I have a Plex pass it is shared out to a couple of people outside of my internal Network. The reason for upgrading the ram I would like to use a VM for one lightweight Linux that will be running all the time, and a couple of Docker containers for things like Pi-hole, Home Assistant, and RTL-SDR applications.



    So let me pick your brain for a moment, knowing what I have and use it for. How would you set it up if it was your server? I have everything backed up off-site and I don't mind redoing everything If necessary.

    My homebrew Nas
    OMV 4
    ASRock760GM-HDV
    AMD Phenom II X6 1065T
    Viper 3 16GB DDR3 1600MHz
    PNY 120GB SSD
    Two: Hitachi 4TB
    In a HP Pavilion case :D

    • Offizieller Beitrag

    I would like to set up all my computers to do rsync remotely to the server.

    If your target is to have backup of the computers on the server, have a look at UrBackup.
    Backing up Windows to OMV


    UrBackup can be used on Linux and Windows clients.

  • If your target is to have backup of the computers on the server, have a look at UrBackup.Backing up Windows to OMV


    UrBackup can be used on Linux and Windows clients.


    Thank you very much macom! Looks like it's fairly straightforward. I installed the plugin. When I have time this evening I will set it up. Off crashtests post
    thanks
    Bruce

    My homebrew Nas
    OMV 4
    ASRock760GM-HDV
    AMD Phenom II X6 1065T
    Viper 3 16GB DDR3 1600MHz
    PNY 120GB SSD
    Two: Hitachi 4TB
    In a HP Pavilion case :D

    • Offizieller Beitrag

    knowing what I have and use it for. How would you set it up if it was your server? I have everything backed up off-site and I don't mind redoing everything If necessary.

    First and foremost, all of this is a matter of opinion. (And opinions vary. :) )
    Second, the risk in trying new approaches is removed with backup. That's another good reason for maintaining multiple backups.
    _________________________________________________________


    I've found, over the last few years, that my current data requirement is around 2TB which fits nicely on a 4TB drive with room to grow. From a user perspective, it's just me and my wife with an occasional guest user. In my case, about 1/2 of the data store contains files I've had long term, that I want to protect from bit-rot. Finally, I backup my LAN clients regularly and want to be able to restore them easily. These factors have shaped my choices in hardware and setup.
    _________________________________________________________


    With the above in mind, I use X86 platforms with a strong preference for server grade hardware and ECC. With that noted, some hardware details and server setup would boil down to the size of the data store.


    - Anything 3TB or smaller I'd run a ZFS mirror (for bit-rot protection). The chassis would have to accomodate at least 3 drives, preferably 4, without add-ons or modification.


    - Anything larger than 3TB would addressed with UnionFS+SNAPRAID. (If the data store is larger than 3TB, I'd assume that the bulk of storage is media files and, with 4K content coming to the forefront, that the data store would probably be growing.) For this purpose, the chassis would need to accomodate 4 drives or more. Otherwise the premium would have to be paid for much larger drives in the +4TB to 10TB range.
    (I'm running a long term test of this scenario, using a collection of drives under 4TB.)


    In either case I'd dedicate one drive running a simple file system like EXT4, in the 3 to 4TB range, for client backup storage, and as a place to store VM's and Dockers. This avoids potential issues with CoW filesystems and UnionFS' use of overlayFS.


    Note that both approaches, ZFS or UnionFS+SNAPRAID, get into additional complexity. Both require maintenance processes and at least casual monitoring, but these things are not difficult to understand and they're easy to automate.
    __________________________________________________________


    The crossover point (3TB) is based on the current cost of hard drives, where the cost per TB starts to climb at 4TB. (There's a premium involved in cost per TB, on average, with drives larger than 4TB.) Of course, with HAMR hard drives on the horizon, the entire calculus may change but I believe it will be awhile before newer storage tech is available and cost effective to consumers.

  • Fractal Design Define R6 Thinking about something like that for a new case something I won't have to upgrade for a long time it's a bit pricey but it's either that or a rack-mount case.


    I've been taking all the DVDs that I've collected since I was a kid and putting them on the NAS and music and surprisingly I got a warning it was 85% full on one of my 4 terabyte drives over the weekend lol. Thank you for all the input time to read up :)

    My homebrew Nas
    OMV 4
    ASRock760GM-HDV
    AMD Phenom II X6 1065T
    Viper 3 16GB DDR3 1600MHz
    PNY 120GB SSD
    Two: Hitachi 4TB
    In a HP Pavilion case :D

Jetzt mitmachen!

Sie haben noch kein Benutzerkonto auf unserer Seite? Registrieren Sie sich kostenlos und nehmen Sie an unserer Community teil!