| To: | Emmanuel Florac <eflorac@xxxxxxxxxxxxxx> |
|---|---|
| Subject: | Re: XFS filesystem reports as full though it isn't |
| From: | David Chinner <dgc@xxxxxxx> |
| Date: | Tue, 20 May 2008 16:10:47 +1000 |
| Cc: | Christoph Hellwig <hch@xxxxxxxxxxxxx>, xfs@xxxxxxxxxxx |
| In-reply-to: | <20080519221057.75e0a775@xxxxxxxxxxxxxx> |
| References: | <20080516222755.3e557c00@xxxxxxxxxxxxxx> <20080518131855.GA1383@xxxxxxxxxxxxx> <20080518132127.GA19006@xxxxxxxxxxxxx> <20080518154208.77b6be3e@xxxxxxxxxxxxxx> <20080519221057.75e0a775@xxxxxxxxxxxxxx> |
| Sender: | xfs-bounce@xxxxxxxxxxx |
| User-agent: | Mutt/1.4.2.1i |
On Mon, May 19, 2008 at 10:10:57PM +0200, Emmanuel Florac wrote:
> Le Sun, 18 May 2008 15:42:08 +0200 vous écriviez:
>
> > IIRC I tried both, however I'll double check monday...
>
> Well obviously I didn't try both, because it worked as expected when
> using "inode64". However there'ssomething slightly weird in the data
> layout:
>
> /dev/dm-0:
> AG Inodes IUsed IFree Use%
> 0 4288 4277 11 99%
> 1 0 0 0 0%
> 2 0 0 0 0%
.....
If there's no free space in AGs > 0 when you enable inode64, then
you will still get enospc...
> See how all inodes are in the same vg? is it OK?
^^ AG
This is typical of inode32 allocation...
Cheers,
Dave.
--
Dave Chinner
Principal Engineer
SGI Australian Software Group
|
| <Prev in Thread] | Current Thread | [Next in Thread> |
|---|---|---|
| ||
| Previous by Date: | [PATCH] always set a/c/mtime through ->setattr, Christoph Hellwig |
|---|---|
| Next by Date: | Re: [PATCH] always set a/c/mtime through ->setattr, Artem Bityutskiy |
| Previous by Thread: | Re: XFS filesystem reports as full though it isn't, Emmanuel Florac |
| Next by Thread: | Re: XFS filesystem reports as full though it isn't, Emmanuel Florac |
| Indexes: | [Date] [Thread] [Top] [All Lists] |