Jump to content

iomgui

Members
  • Posts

    14
  • Joined

  • Last visited

iomgui's Achievements

Member

Member (2/3)

0

Reputation

  1. Hello, I tried many ways but I cant find any list of changelog for the different versions of drivepool. Is it available somewhere ? Many thanks
  2. The only easy solution i found (easy = not requiring that you intevrene each time you add a folder) is to set balancers so that they fill one disk to 90% (for example) before starting to fill another disk. then you can either : - do nothing : only 1 folder could have files on 2 disks when it changes disk to the next one, when one disk reached max capacity as defined by you - or you could every now and then (every 3 months for example), change the rule to fill another disk even if you have reached only 30 or 40% for example. That way you will have 0 folder with files on more than 1 drive But i agree, such a balancer option would be nice, but mostly concerns people without parity disks or duplicate (managed by drivepool). If you have a backup that is not at the same place (only strategy to prevent complete lost in case of fire or flooding, or burglary), then you can see how usefull it would be
  3. This one illustrates it as usefull to store movies with limiting the number of disk spining, and keeping balance between disks. I share the same need, with also the need that when one disk fail, i accept to loose some folders, but it would be a problem to loose part of all my folders because of the way i archive things, the amount of work to rebuild the data without duplication would be HUGE (find the files missing and rescanning them not by batch but by picking one or 2 files in any physical folder cold storage from the office, for example). I know i could duplicate or use snapraid, but even with snapraid for example, with 1 parity disk, you still have to consider a 2 disk failure. And in the case of a 2 disk failure i need to minimize the effort of rebuilding the data. But its very similar to the movie collection explained in this post. https://community.covecube.com/index.php?/topic/5889-is-it-possible-to-craft-file-placement-rules-such-that-files-in-folders-stay-grouped-together/
  4. I understand, but it wont do what i need : for each new folder created on the level where no disk separation should be done, it will not be possible to automaticly create this folder and its content on one and only one aleatory disk, choice of alleatory following other balance rules. (it would work if the directory where its created in is already assigned a physical drive, but its not the case when you create it at the level where the choice of physical drive should be allowed). Because the simpliest solution would then be to only allow to fill one disk with existing plugins, withouth the need to create rules at directory level, and then switch after a few months to another physical disk (without any plugin rebalancing depending on disk equilibrium). By simpliest solution, i mean the one that would require the minimum number of manual intervention on my side to change balancing rules or create directory rules. Though this simpliest solution has the huge disadvantage to not share the risk of disk failure by balkancing the amount of data between disks. I dont see any way to automate it so far, and i found a post in the meantime saying it was not possible with actual pluggins. I can link it here if its authorized and usefull
  5. Thanks, But i dont see how for incoming folders that would be created on the 3rd levels. For existing folders i can select, one by one, to balance them on only one drive. But i m looking for a way to do this automaticly so i dont have to add a rule for any 3rd level folder that would be created and probably with a content already dispatched on different physical drives. Unless i missed something
  6. Hello everyone Ok here is what i would like to do, lets say my structure on the pool is like this : folder A / sub Ai / Sub Sub Ai,j folder B / sub Bi / Sub Sub Bi,j etc At some point in folder hierarchy, lets say starting at the 3rd subfolder (for exemplae sub sub A i,j). What i would like is a way when everytime i transfer a folder that is at 3rd or more deeper in foilder structure, that this folder content is always put on the same drive, as long as there is enough space. I do not care on which drive it is put, just that it is maintained on only one drive. Also, anytime i m adding files to that folder, thoses files should follow the same rule and be put on the same drive as the drive of the original folder. I m looking for a rule that could be set once and would never require me to a destination drive. I m also wondering if such a rule could rebalance an existing pool rearranging every files so that SubSubA etc are only on one physical drive, even though its not the case actually because of what driveopool did so far. Many thanks in advance
  7. ECC memory is supposed to be able to autocorrect on the fly 1 bit of corrupted data in memory. With ECC then only HDD corruption could be considered (i do not consider corruption from cpu). For only HDD data corruption, then snapraid does answer to the problem as long as it s a per request check. But snapraid doesnt do it every time the file is opened for example, only when the bacth scan will be launched (though its enough in my case). I was more wondering if scanner would be of any benefit if not using ECC memory compared to the case where you could use ECC memory ... Or another way to say it : if scanner does allow me to use non ECC memory with a result as good as if i used ECC memory without scanner, then i would be interested because chossing ECC memory does limt a lot the choice of other components … unless price doesnt matter
  8. ECC memory is supposed to be able to autocorrect on the fly 1 bit of corrupted data in memory. With ECC then only HDD corruption could be considered (i do not consider corruption from cpu). For only HDD data corruption, then snapraid does answer to the problem as long as it s a per request check. But snapraid doesnt do it every time the file is opened for example, only when the bacth scan will be launched (though its enough in my case). I was more wondering if scanner would be of any benefit if not using ECC memory compared to the case where you could use ECC memory ...
  9. Lets say i m using drivepool for data only, and not on the drive that is used for the OS installation. Now lets suppose i need to change both the motherboard and the drive on which the OS is installed, without being able to preprare anything (HDD of OS + motherboard faulty). Once the new motherboard and OS is installed, what would be (if any) the way to recover the "old" ? is possible without the need of balancing any data between the disks that were incolved in the pool, in otherwords, with every files remaining on the drives where they actually are
  10. Considering i will use snapraid and update parity drive + checksums everry night. In what scenario would Scanner be of benefit to me if : 1/ i am also using ECC memory 2/ without using ECC memory By scenario, i mean in what practical cases it would recover from or prevent corrupted date ? (i just want to measure the possible benefit before adding a layer to my NAS media center). Also : when does scanner check the parity ? anytime the file is open ? and or / scheduled ? and or ?
  11. Hello, I m trying it right now after being disappointed by space storages lack of easy customization. I will use snapraid along with drivepool, for info. My concern is more about the project being still in developpement or getting abandonned progressivly by its author, as i have seen very few communications (not to say none …) about incoming features or bugs being fixed / improvements.
×
×
  • Create New...