Darksurf

Members
  • Posts

    166
  • Joined

  • Last visited

Posts posted by Darksurf

  1. HellDiverUK, I'd thought about that myself. But I checked the BTRFS website (as well as several other angry sources) and its a known issue and claims to be fixed in 3.19.5 and 4.x Going forward. So this exact issue is INDEED a BTRFS issue. RAID isn't Hardware raid either, its BTRFS software RAID. Until I see it proven solid, the cake is a lie.

  2. I've not used BTRFS in unRAID yet. Ordered Unraid 6 on a USB 3 drive not to long ago and its in the mail. But I did test BTRFS on the 3.19.3 kernel on my Sabayon tower and I can report it is DISASTROUS! I made a RAID10 of 4 1TB HDDs (WD REDS) in a new ICY DOCK USB3 enclosure. All went well at first and I began debating ditching ZFS for a more Linux friendly/supported FS. So I started to migrate (copy) all my data (~1.8TB) off my FreeNAS VM through the network to this BTRFS RAID. Worked great for several hours and was fairly speedy. Right before the 24 hr marker I noticed all my files had "finished" being copied over. I went to double check only to find about 1TB of data and the drive had been marked RO? I looked in dmesg and found tons of complaints about metadata corruption on "sdg". Looked into it, this is a known common issue until kernel 3.19.5 and there IS NO FIX. I would have to remove the drive from the RAID, wipe it, then re-add it to the RAID. SMART data shows drive is fine and all works fine with ZFS.

     

    Sadly BTRFS is garbage and instable in my eyes. If I lose a drive in RAID10 due to the FS in less than 24hrs, forget that crap! I'd much rather have ZoL.  Devs who claimed businesses should use this FS as "it is stable now" are a bunch of liars and are full of it. Wait for 4.0 kernel or.. wait until someone else takes a dive and it works. I'll probably give it another try when 4.0 comes out and if I see good reviews.

     

    Please devs, include ZFS support with the 4.0 kernel. I like my data and prefer to keep it.