New system, blk_update_request: I/O error, dev xxx, sector xxxxxxxx on all 8 drives

  • HI OMV Forum,


    Completely new to OMV and linux, hoping someone out there knows something about how to fix the issues i'm having as i'm at my wits end. I know nothing about command based OS' so not the best time to encounter such a serious issue when i'm just getting started with it ; (
    I don't know if there is any system info, log, screenshots etc. you need more than what i've attached, but please let me know and how to get it and it will be yours ;D


    Just switched out a QNAP (wildly underpowered non-upgradeable Celeron CPU) to get a bit more serious with things and bought an old Dell T610 server, 16GB ram with dual Xeon E5506 CPU's (in case you are wondering why such old hardware, very very cheap 8 bay system with 2 extra potential bays in the 2x redundant 5.25 bays with a bracket for 3.5" HDD.


    Since I knew it had an old PERC 6/i RAID controller that can't support bigger drives than 2TB I purchased and installed an LSI-9211-8i Raid/HBA controller (might have been an IBM M1015 originally, was already flashed to IT mode from seller so I didn't end up bricking it).
    Also purchased 2 x Xeon X5675 CPU's for 25$ each to give it enough computing power for transcoding 3-4+ simultaneous 1080p streams as I only intend to use the system for media/Plex server. After CPU switch went from 7000 to 19.000 score in geekbench, which is all I need so so far so good.


    OMV installed on 16GB Sandisk cruzer Fit with Flashmemory plugin in OMV (although I hear SSD for OS is recommended but don't think this issue is related). Extra Cruzer Fit bought for backup of OS in case of burn out but not yet configured to do so.
    1 x SSD connected directly to the motherboard via SATA used for metadata storage, plex, transcoding etc.
    As i'm not running RAID or any other enterprise applications googling seemed to suggest that EXT4 is the filesystem to choose, for what is essentially just mainly basic media file storage.


    Basically the problem is that right off a new OMV install it gives blk_update_request: I/O error, dev xxx, sector xxxxxxxx on all drives (4 x 8TB WD RED and 4x 10TB WD RED)
    When drives weren't mounted it gives the errors right away when accessing the web interface of OMV. When drives are mounted that stopped happening and I thought the issue was a non issue as long as they were mounted, I could live with that and contributed it to some minor incomparability between the LSI card and either western digital drives or 8TB/10TB drives in general, it is an old card after all, even though it detects them fine 8TB and 10TB drives.


    But as I started copying data to the mounted drives the I/O errors started happening again, only once did it seem to stop the transfer to the drive. When I check the byte size between the copied data there are no differences and things play nicely so it doesn't seem to corrupt the files.
    Also when playing a video file from the drives it results in a lot of I/O errors but video plays fine. However when I try to unrar something located on one of the drives it doesn't seem to like that and the filesystem on that drive crashes almost every time and reboots in read-only. As more and more data gets put on the harddrives the errors are becoming more frequent. Also sometimes while rebooting it will do the systemd-fsck check of drives but get interrupted because of IO errors and they say there is something wrong with the file system. Haven't been able to capture screenshot of it yet.


    The LSI-9211-8i is connected to the Dell backplane/hot swap drive bays with 2 x mini sas SFF-8087 cables which I obviously tried to replace to see if cables we're damaged but same result.
    When googling this issue I haven't been able to find anyone with exactly the same problem that had a HBA/HDD setup resembling mine.. Closest seemed to be someone who had the same I/O errors where it got solved by switching from IDE mode to AHCI in bios but as far as I can understand that's only relevant on a pure SATA connection and I can find nowhere in BIOS or controller card settings that lists IDE/AHCI mode. Maybe i'm overlooking something? Could the motherboard itself be incompatible somehow?


    Most common answer to an I/O error problem seems to be that harddrive is about to die but obviously with 8 fairly new drives I doubt it's that. Also there were some mention online of I/O errors could be a power thing, the server has 2 power supplies and when it reboots it gives some ACPI errors as well as something about an array and 1 drive out of 32 but i'm not running raid or at least not supposed to, don't know how to look further into any of that and didn't think it was related. Everything in attachments to this post


    My 2 remaining ideas are that its either an issue/defect/incompatibility with the original Dell Backplane or the LSI-9211-8i and these large hard drives which didn't exist when they we're made. When i've tried with older Hitachi 1TB-2TB drives I haven't seen (at least noticed) any IO errors.
    Apart from ripping out my remaining hair.. what I was going to try next was to purchase a newer controller card to test if it's the LSI 9211-8i that's incompatible/defective and to buy 2 x SFF-8087 mini sas to 4x SATA breakout cables to bypass the backplane and thereby testing if thats where the issue lies.
    However as I don't live in US/EUR this hardware is not locally available and Ebay/Amazon is slow and expensive so was hoping to exhaust every other possible solution before ordering parts I potentially don't need.


    Am I on the right track here or am I overlooking some obvious things with the EXT4 filesystem, OMV, Debian, HBA controller setup, etc all of which are completely new to me.
    Hope some friendly knowledgeable folk out there can help, greatly appreciated.

  • Sorry, but your post is a little bit too long. I am ?( Could you please describe your problem in a nutshell ;)

    OMV 3.0.100 (Gray style)

    ASRock Rack C2550D4I C0-stepping - 16GB ECC - 6x WD RED 3TB (ZFS 2x3 Striped RaidZ1) - Fractal Design Node 304 -

    3x WD80EMAZ Snapraid / MergerFS-pool via eSATA - 4-Bay ICYCube MB561U3S-4S with fan-mod

  • hey there, sorry for the length but felt it important to describe in detail


    in a nutshell the issue is that i'm a first time OMV user, just installed it and a new HBA and it gives me errors on all drives
    blk_update_request: I/O error, dev sdd, sector 339393

  • Could you connect some of the drives directly to the mainboard in order to check without the HBA controller and without using BTRFS? Keep it simple at the beginning :) Did you have wiped the drives?

    OMV 3.0.100 (Gray style)

    ASRock Rack C2550D4I C0-stepping - 16GB ECC - 6x WD RED 3TB (ZFS 2x3 Striped RaidZ1) - Fractal Design Node 304 -

    3x WD80EMAZ Snapraid / MergerFS-pool via eSATA - 4-Bay ICYCube MB561U3S-4S with fan-mod

  • directly to mainboard works but again only 2TB support
    Instead of fiddling more with it I purchased a newer 9207-8i card instead and works like a charm
    thank you anyway

Jetzt mitmachen!

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