Mail Archive Search: <+subject:/^(?:^\s*(re|sv|fwd|fw)[\[\]\d]*[:>-]+\s*)*Pathological\s+allocation\s+pattern\s+with\s+direct\s+IO\s*$/>

Results:

References: [ +subject:/^(?:^\s*(re|sv|fwd|fw)[\[\]\d]*[:>-]+\s*)*Pathological\s+allocation\s+pattern\s+with\s+direct\s+IO\s*$/: 9 ]

Total 9 documents matching your query.

1. Pathological allocation pattern with direct IO (score: 1): Author: Jan Kara <jack@xxxxxxx>; Date: Wed, 6 Mar 2013 21:22:10 +0100; Hello, one of our customers has application that write large (tens of GB) files using direct IO done in 16 MB chunks. They keep the fs around 80% full deleting oldest files when they need to store ne; /archives/xfs/2013-03/msg00144.html (8,994 bytes)
2. Re: Pathological allocation pattern with direct IO (score: 1): Author: Ben Myers <bpm@xxxxxxx>; Date: Wed, 6 Mar 2013 16:01:55 -0600; Hi Jan, I believe we've seen something similar to #2 before: /data/dbench.dat: EXT: FILE-OFFSET BLOCK-RANGE AG AG-OFFSET TOTAL FLAGS 0: [0..150994943]: 2343559168..2494554111 5 (2048..150996991) 1509; /archives/xfs/2013-03/msg00146.html (12,920 bytes)
3. Re: Pathological allocation pattern with direct IO (score: 1): Author: Dave Chinner <david@xxxxxxxxxxxxx>; Date: Thu, 7 Mar 2013 16:03:25 +1100; Well known issue, first diagnosed about 15 years ago, IIRC. Simple solution: use extent size hints. Cheers, Dave. -- Dave Chinner david@xxxxxxxxxxxxx; /archives/xfs/2013-03/msg00155.html (10,316 bytes)
4. Re: Pathological allocation pattern with direct IO (score: 1): Author: Jan Kara <jack@xxxxxxx>; Date: Thu, 7 Mar 2013 11:24:06 +0100; I thought someone must have hit it before. But I wasn't successful in googling... I suggested using fallocate to the customer since they have a good idea of the final file size in advance and in test; /archives/xfs/2013-03/msg00158.html (12,762 bytes)
5. Re: Pathological allocation pattern with direct IO (score: 1): Author: Mark Tinguely <tinguely@xxxxxxx>; Date: Thu, 07 Mar 2013 07:58:35 -0600; On 03/07/13 04:24, Jan Kara wrote: On Thu 07-03-13 16:03:25, Dave Chinner wrote: On Wed, Mar 06, 2013 at 09:22:10PM +0100, Jan Kara wrote: Hello, one of our customers has application that write large; /archives/xfs/2013-03/msg00170.html (13,001 bytes)
6. Re: Pathological allocation pattern with direct IO (score: 1): Author: pg_xf2@xxxxxxxxxxxxxxxxxx (Peter Grandi); Date: Wed, 6 Mar 2013 22:31:42 +0000; They just want *guaranteed* (not just nearly-always) contiguity, without preallocating, with incremental writes with direct IO, while keeping the filesystem full at 80%. And a pony! :-); /archives/xfs/2013-03/msg00185.html (8,862 bytes)
7. Re: Pathological allocation pattern with direct IO (score: 1): Author: Dave Chinner <david@xxxxxxxxxxxxx>; Date: Fri, 8 Mar 2013 12:35:25 +1100; ..... Extent size hints have the advantage of being effective when the file size is not known ahead of time, or the application cannot be modified to do preallocation. And they can be applied immedia; /archives/xfs/2013-03/msg00194.html (19,334 bytes)
8. Re: Pathological allocation pattern with direct IO (score: 1): Author: Jan Kara <jack@xxxxxxx>; Date: Tue, 12 Mar 2013 12:01:00 +0100; <snip> Thanks for explanation! Yup, that's what I understood in the code and saw in my experiments... Yes, I was wondering about whether this strategy won't be good enough as well. But there's one ca; /archives/xfs/2013-03/msg00286.html (17,135 bytes)
9. Re: Pathological allocation pattern with direct IO (score: 1): Author: Dave Chinner <david@xxxxxxxxxxxxx>; Date: Fri, 15 Mar 2013 10:36:31 +1100; No problems - there's a lot of stuff in my head that simply isn't documented anywhere, and it would take months of work to document it so I've never written it down... It took me a couple of readings; /archives/xfs/2013-03/msg00432.html (15,801 bytes)

Current List: 1 - 9
Page: [1]