[Gluster-devel] Re: Unify - writev reschedule file after ENOSPC (No space left on device)

[Tomáš Siegl]

Hello

Probably my question was lost in time :-).

I would like to know the opinion of the gluster team if you plan to modify
Unify behaviour in future to support this case?

Best regards
Tomas Siegl

On Mon, Sep 29, 2008 at 3:25 PM, Tomáš Siegl <tomas.siegl at gmail.com> wrote:

> Hello
>
> I have question do you plan to extend Unify translator for functionality
> that
> if during writing to file is all available space on child node exhausted
> unify
> translator will reschedule file to another child node?
>
> E.g. Asume that I have 3 brick with 10, 20, 30 GB Free space Unified into 1
> disk
> space (max 60 Gig Free space). Now I would like to upload 15GB file. Unify
> will schedule file for server with 10G space. During upload I will get
> error message
>
> dd: closing output file `test': No space left on device
>
> This test was done using NUFA scheduler and verified from latest 3.0
> tla source codes.
>
> Regards
> Tomas Siegl
>
> IRC #Gluster nick: SigiTS
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://supercolony.gluster.org/pipermail/gluster-devel/attachments/20081022/6cbfa350/attachment-0003.html>