Search String: Display: Description: Sort:

Results:

References: [ +subject:/^(?:^\s*(re|sv|fwd|fw)[\[\]\d]*[:>-]+\s*)*Deleting\s+files\s+with\s+extended\s+attributes\s+is\s+dead\s+slow\s*$/: 9 ]

Total 9 documents matching your query.

1. Deleting files with extended attributes is dead slow (score: 1)
Author: Bernd Schubert <bernd.schubert@xxxxxxxxxxxxxxxxxx>
Date: Fri, 12 Aug 2011 20:19:53 +0200
Hi there, with FhGFS we may use extended attributes to store file meta data and with ext3/ext4 that also works very well and the rate to create files and to write those EAs (create() + fsetxattr() is
/archives/xfs/2011-08/msg00233.html (10,960 bytes)

2. Re: Deleting files with extended attributes is dead slow (score: 1)
Author: Christoph Hellwig <hch@xxxxxxxxxxxxx>
Date: Fri, 12 Aug 2011 16:47:46 -0400
Currently XFS uses synchronous transactions when retiring inodes with xattrs on them. So far I've not actually found a good reason for that, so I suspect we can remove it. I'll do another audit of it
/archives/xfs/2011-08/msg00240.html (8,093 bytes)

3. Re: Deleting files with extended attributes is dead slow (score: 1)
Author: Christoph Hellwig <hch@xxxxxxxxxxxxx>
Date: Tue, 16 Aug 2011 12:13:58 -0400
The patch below should help with the performance of deleting files that have external attributes. But I still wonder why you see attributes outside the inode if you are using 512 byte inodes and arou
/archives/xfs/2011-08/msg00293.html (10,955 bytes)

4. Re: Deleting files with extended attributes is dead slow (score: 1)
Author: Bernd Schubert <bernd.schubert@xxxxxxxxxxxxxxxxxx>
Date: Wed, 17 Aug 2011 15:05:28 +0200
On 08/16/2011 06:13 PM, Christoph Hellwig wrote: The patch below should help with the performance of deleting files that have external attributes. But I still wonder why you see attributes outside th
/archives/xfs/2011-08/msg00302.html (15,276 bytes)

5. Re: Deleting files with extended attributes is dead slow (score: 1)
Author: Christoph Hellwig <hch@xxxxxxxxxxxxx>
Date: Wed, 17 Aug 2011 13:02:52 -0400
It means that you are having an extent spanning 8 blocks for xattr storage, that map to physical blocks 92304 to 9231 in the filesystem. It sounds to me like your workload has a lot more than 256 byt
/archives/xfs/2011-08/msg00304.html (10,068 bytes)

6. Re: Deleting files with extended attributes is dead slow (score: 1)
Author: Bernd Schubert <bernd.schubert@xxxxxxxxxxxxxxxxxx>
Date: Wed, 17 Aug 2011 19:39:01 +0200
On 08/17/2011 07:02 PM, Christoph Hellwig wrote: On Wed, Aug 17, 2011 at 03:05:28PM +0200, Bernd Schubert wrote: (squeeze-x86_64)fslab2:~# xfs_bmap -a /mnt/xfs/Bonnie.29243/00000/00000027faJxifNb0n /
/archives/xfs/2011-08/msg00305.html (13,325 bytes)

7. Re: Deleting files with extended attributes is dead slow (score: 1)
Author: Dave Chinner <david@xxxxxxxxxxxxx>
Date: Thu, 18 Aug 2011 12:08:48 +1000
XFS directory algorithms are significantly more complex than ext4. They trade off CPU usage for significantly better layout and scalability at large sizes. i.e. CPU costs less than IO so we burn more
/archives/xfs/2011-08/msg00307.html (13,207 bytes)

8. Re: Deleting files with extended attributes is dead slow (score: 1)
Author: Dave Chinner <david@xxxxxxxxxxxxx>
Date: Thu, 18 Aug 2011 13:05:19 +1000
It is neither of these. Ok, I've reproduced it on a 3.0-rc2 kernel with attr=2 and 2k inodes. An attribute of 254 bytes stays in line: .... u = (empty) a.sfattr.hdr.totsize = 270 a.sfattr.hdr.count =
/archives/xfs/2011-08/msg00308.html (15,411 bytes)

9. Re: Deleting files with extended attributes is dead slow (score: 1)
Author: Christoph Hellwig <hch@xxxxxxxxxxxxx>
Date: Mon, 22 Aug 2011 10:35:29 -0400
Btw, one interesting thing I noticed is that xfs_attr_leaf_create always zeroes and logs whe whole attr block. I'll need to audit this a bit more, but from the highlevel POV I don't see why we can't
/archives/xfs/2011-08/msg00360.html (9,688 bytes)


This search system is powered by Namazu