Search Results

Search results 1-20 of 119.

  • Mergerfs wakes harddisc

    trapexit - - General

    Post

    You wouldn't concat the drives together. You'd create for each drive their own physical and logical volume and put the filesystem on top of that. Then partition the ssd N ways and give one cache and metadata partition to each.

  • Mergerfs wakes harddisc

    trapexit - - General

    Post

    Really the only way would be to run it manually in debug mode. `-d` Ideally you'd disable whatever is querying the data. The OS caches some info but if the OS asks mergerfs for data it has no choice but to passthrough the request. IMO adding caching to mergerfs isn't the solution. There are already multiple caches in the stack. I've not yet tried it but one idea is to use lvm caching in front of your hard drives but that'd require using lvm on all target drives. If you don't need to point those …

  • Thanks. I've been wanting to rewrite it to have a nicer workflow. Currently it just starts moving files around. Preferably it'd calculate what should be moved and where and then provide a summary before acting on it. But... it works, I don't use it that often, and I've been working on other things.

  • Policies are used at the point the request is made. mergerfs doesn't actively move data around and that's very purposeful. Best to leave such complicated tasks to the user. If your drives are filling in ways you don't want its likely you chose the wrong policy. But tools like mergerfs.balance can help but should only be used if you don't care where data goes as I never made its algo any more complicated than move files from the fullest drives to the least fullest.

  • Mergerfs wakes harddisc

    trapexit - - General

    Post

    I'm not entirely sure what you're asking. mergerfs doesn't store anything for itself on the drives and doesn't query the underlying drives on its own. The access pattern depends on something requesting it and the policy of the request.

  • Mergerfs wakes harddisc

    trapexit - - General

    Post

    It's not mergerfs. Something is using the filesystem. Many media services regularly query / use the filesystem. Go through different services you have running and disable them.

  • The order of options means nothing unless you're duplicating an option (last version wins). I'm not sure what you're asking exactly. If you wish to store Plex's config through mergerfs you have to disable direct_io and as mentioned in the docs it is suggested you turn on drop cache which you have. Ideally Plex would not use mmap but it is what it is. If you put the config elsewhere you can use direct_io just fine. My own library is available to plex through a direct_io enabled mount.

  • mergerfs.balance doesn't work

    trapexit - - Plugins

    Post

    You downloaded a website. Not the code. If you're going to download the file directly you have to click on "raw". wget raw.githubusercontent.com/trap…4c87/src/mergerfs.balance Also if you download the zip file or clone the repo there is a "make install" step which would install all to tools into /usr/local/bin

  • Suggestion on the set up. RAID or?

    trapexit - - RAID

    Post

    This has little to do with OMV. What precisely is the behavior you are looking for?

  • mergerfs.balance doesn't work

    trapexit - - Plugins

    Post

    The balance tool does not care about existing paths. If you care about path preservation then you shouldn't use mergerfs.balance. That said... by using the exclude path option you could exclude specific drives in the pool from being targets so you could balance files to a subset of drives. It might not be too much work to add a flag to filter targets by available directory but I'm not sure it'd work too well TBH. It'd be cool to make a tool which would try to keep data locality while also balanc…

  • mergerfs.balance doesn't work

    trapexit - - Plugins

    Post

    I'm guessing you aren't too familiar with Linux command lines. If you don't put those arguments in quotes the shell expands the asterisk rather than passing it into the app. $ sudo mergerfs.balance -s 10M -e "*aquota*" -e "snapraid*" -E "*docker/*" -E "*downloads/*" /source/directory

  • mergerfs.balance doesn't work

    trapexit - - Plugins

    Post

    It's an fnmatch pattern just like the others. It just checks the whole path rather than file. That path is the full source path, not the mergerfs path. The real intent is for you to provide relative path matches. "*/docker/*" or similar. Are you wrapping those arguments in quotes? It doesn't print anything out if it's scanning.

  • mergerfs.balance doesn't work

    trapexit - - Plugins

    Post

    OK... try the latest version: github.com/trapexit/mergerfs-t…ster/src/mergerfs.balance Has path and size excluding / including. Also works on subdirectories.

  • mergerfs.balance doesn't work

    trapexit - - Plugins

    Post

    Sorry for the delay. I've got something working just need to run more tests. Also adding some tweaks such as being able to act not just at the mount level but in subdirectories.

  • mergerfs.balance doesn't work

    trapexit - - Plugins

    Post

    Sure. I'll add it.

  • mergerfs.balance doesn't work

    trapexit - - Plugins

    Post

    The exclusion works on filenames. Not paths. The tool would need to be reworked slightly to use full directory paths. edit: It's a pretty straight forward change. I'll see if I can work on it today or tomorrow.

  • While in the end this had nothing to do with mergerfs... it's always a good idea when testing such things to start at the bottom and move up. Do some basic dd based tests on the native drive, then mergerfs, then locally through the network filesystem then remotely. mergerfs actually has a special mode for testing. The `nullrw` option causes mergerfs not to actually read or write to the backing devices which gives you theoretical max throughput. You can check the mergerfs docs for details.

  • It currently only works at the entire pool level. Haven't gotten around to doing single sub directories yet.

  • If you don't care what data lands where then the balance script can move data around for you.

  • mergerFS add disk

    trapexit - - General

    Post

    No, you don't need to create directories manually if you aren't using path preserving. The policies should work exactly as described in the docs [0]. Are the permissions correct on the new drive? Is it the drive with the most free space? [0] github.com/trapexit/mergerfs#functions--policies--categories