Jump to content

Scintilla13

Members
  • Content Count

    8
  • Joined

  • Last visited

Community Reputation

1 Neutral

About Scintilla13

  • Rank
    Newbie
  1. I'm still not completely happy with it, anyway I opted for OpenMediaVault 5. I Installed on it MergeFS and SnapRaid, basically is the same solution without many whistles. OMV is based on Debian and it is more open to any need, if you need a packet you are free to install. I have fear of "closed" solutions like unraid, is not a matter of price. I'm also not happy of what I found on this forum. I think Unraid is a great work but more suitable for an home media center. Different uses are seen with diffidence.
  2. Because here there is a lot of people complaining about parity disk rebuilding process. Actually I don't need to parity protect data like VMs backups.
  3. I recently started using UnRaid as a Proxmox network storage. UnRaid is more highly scalable than any other NAS solution. Ceph requires a too much expensive minimal hardware setup. ZFS is the most awful and slower thing ever (I know... but this is in very short my personal opinion and experience). I'm running 10Gbps Ethernet on UnRaid and Proxmox hosts. The UnRaid PC has 32GB RAM and 9gen six core i5. NFS is out of the way because it doesn't reconnect by itself so I switched to SMB. To reach my goals I would like to get 3 different "levels" of storage, I'm going to summarize their scope and then I will explain my consideration about my unRaid knowledge. a) SSD storage for performance demanding VMs b) HDD storage for regulars VMs c) storage "without care" for VMs backup a) I created a unRaidSSD share then I could use with "cache = only" or "cache = prefer" settings. The fact that some data exists ONLY in the SSD without being also present in the parity protected array worries me a lot. In my idea there should be not only a mover, but a background sync process between SSD cache and HDD (rsync?). it could be a new feature request or is it just an impossible to implement perversion in my mind? b) unRaid parity protected array share with "cache = no" setting During long I/O Mover operation I got some VM data corruption. Reading this forum, I'm achieving that I will get the same when I will add a parity drive. There is a way to set Mover and parity rebuild process as background priority? Because VMs are made by a single huge file they will not benefit of any "cache = yes" setting, except occasional snapshots. c) unassigned drive share without any parity In general the possibility to manage a parity excluded array could be a bonus. What are your advice and considerations (including unRaid is not the right solution) ? Thanks!
  4. The request starts from a need: I have a lot of backup data that really I don't care. For what is my current unRaid knowledge unassigned disks are going to be unsupported. Unassigned disk are not featuring smart array features as data spanning over multiple HDD. I don't know if is best to implement it as: - put all drives in the array and exclude some from parity - manage a non parity array Or maybe this already exists and I am a total newbye?!?! Thanks!
  5. +1 to allow parity buildup process priority to be editable.
  6. Thank you! I will post any difficulty on the other thread.
  7. I have the same question, but in this post there is no link to a solution neither which configuration implement. Please could you help me on how to configure a BTRFS JBOD cache volume with replaceable devices?