Howdy, Stranger!

It looks like you're new here. If you want to get involved, click one of these buttons!


Are there any tried and tested script to mount Hubic in the file system?
New on LowEndTalk? Please Register and read our Community Rules.

All new Registrations are manually reviewed and approved, so a short delay after registration may occur before your account becomes active.

Are there any tried and tested script to mount Hubic in the file system?

Are there any tried and tested script to mount Hubic as a read/write disk in the file system?

Something like the S3/FUSE clients available for Amazon?

Comments

  • OliverOliver Member, Host Rep
    Thanked by 1rchurch
  • yes, but very slow

    Thanked by 1rchurch
  • I gave up on Hubic due to speed. Love the idea of 10Tb for €10, but the service and software suck.

    Thanked by 1rchurch
  • @ndlong75, @squibs

    Do you consider the slow speed due to the inherent nature of FUSE software itself, or was it down to the network speed ?

    A recent post indicates that Hubic's network speed has greatly increased.

  • rm_rm_ IPv6 Advocate, Veteran

    hubicfuse works, as for speed, couldn't you try it yourself on the free account?
    The free 25GB mounted into the FS are enough for a lot of purposes.
    And no, FUSE will not be a bottleneck if we're talking about a service accessed over the Internet.

  • Does a Hubic account only work on a single computer, or can you link it with multiple devices?

  • @rm_ said:
    hubicfuse works, as for speed, couldn't you try it yourself on the free account?
    The free 25GB mounted into the FS are enough for a lot of purposes.
    And no, FUSE will not be a bottleneck if we're talking about a service accessed over the Internet.

    Well... FUSE has some serious overheads sometimes, like terrible buffer management.

  • rm_rm_ IPv6 Advocate, Veteran

    rchurch said: can you link it with multiple devices?

    Yes.

  • I found FUSE slow when working with Box.com, uploading over HTTP from a browser ended up being faster...

  • I mostly tested it on a windows box, with Hubic's own client and with 2 third party clients. I'm not sure about the underlying tech, but they were all really slow.

  • squibssquibs Member
    edited September 2014

    Update - one of the clients I tried used webdav and was painfully slow. To put this in perspective, I have a webdav connection to owncloud running on a kimsufi server, and that's an order of magnitude quicker, so the speed problem is with Hubic.

    I just cancelled my account.

  • rm_rm_ IPv6 Advocate, Veteran

    squibs said: Update - one of the clients I tried used webdav and was painfully slow. To put this in perspective, I have a webdav connection to owncloud running on a kimsufi server, and that's an order of magnitude quicker, so the speed problem is with Hubic.

    That's certainly "very interesting", considering Hubic does not even support WebDAV in the first place, so it's far from being clear what did you test and what's the relation to this thread.

  • @rm_ said:
    That's certainly "very interesting", considering Hubic does not even support WebDAV in the first place, so it's far from being clear what did you test and what's the relation to this thread.

    I was using expandrive to map Hubic as a network drive - my understanding was that it was using webdav under the hood - expandrive does support webdav. As for the relation to the thread, see the thread title.

  • Hubicfuse works well enough but it insists on making use of /tmp unless you hack about with it. There is also another branch that supports large files. Essentially hubicfuse is a version of cloudfuse with the hubic authentication. Hubic mostly implements the openstack swift API but authentication is different. Using hubicfuse on an ovh box I get pretty good speeds.

  • rm_rm_ IPv6 Advocate, Veteran
    edited September 2014

    abat said: There is also another branch that supports large files.

    And that one doesn't use /tmp? Also what's considered "large"? So far I'm only uploading about ~600MB-sized archives there, works fine.

  • openstack swift limits to 5GB after which you have to jump through hoops and upload in parts with a manifest to put them back together. I still had to recompile the DLO branch to get rid of tmpfile() though as I kept running out of disk space in the /tmp partition with large files.

  • rm_rm_ IPv6 Advocate, Veteran

    abat said: I still had to recompile the DLO branch to get rid of tmpfile() though

    If you patched it, can you post the patch somewhere, or better yet, also send to the original author. I'd love it to stop using /tmp, as I have /tmp mounted as tmpfs on most computers.

  • abatabat Member
    edited September 2014

    The replacement is essentially this:

    // FILE *temp_file = tmpfile();
    const char *dir = "/home/tmp";
    const char *pfx = "file_XXXXXX";
    char *temp_file_name = tempnam("/home/tmp", "file_XXXXXX");
    FILE *temp_file = fopen(temp_file_name,"w+b");

    ... rest of function here

    //need to close the file as tmpfile() created auto-remove and we haven't
    remove(temp_file_name);

    I do still have an issue reading large files as they still fill up /tmp and I can't find a tmpfile() I haven't replaced.

Sign In or Register to comment.