* [Ocfs2-devel] du_enhancement v3
@ 2010-03-08 15:30 jeff.liu
2010-03-08 15:44 ` jeff.liu
0 siblings, 1 reply; 2+ messages in thread
From: jeff.liu @ 2010-03-08 15:30 UTC (permalink / raw)
To: ocfs2-devel
Hello,
The following is the v3 patchset for du enhancement.
The bugs fixed according to Tao's comments.
For now, I faced with another problem need your guys help confirm:
I skipped the inline file processing which its data mixed with metadata.
IMHO, it has pretty minor effect for the precise measurement of footprint, as a result,
the rebree process overhead is unworthy for inline file.
Any thoughts or suggestions for this case?
I run a few simple testing, it works:
>>> 1542 - 1 - 21 - 99 - 512 - 105 - 5
799
jeff at jeff-laptop:/shared-du -E -m -c /ocfs2/*
0 /ocfs2/lost+found
1 /ocfs2/reflink
1 (1) /ocfs2/test1
1 (1) /ocfs2/test1r
21 (21) /ocfs2/test2
21 (21) /ocfs2/test2r
99 (99) /ocfs2/test3
99 (99) /ocfs2/test3r
40 /ocfs2/test4
22 (22) /ocfs2/test4r
512 (512) /ocfs2/test5
512 (512) /ocfs2/test5r
105 (105) /ocfs2/test6
105 (105) /ocfs2/test6r
0 /ocfs2/test7
0 /ocfs2/test7r
5 (5) /ocfs2/test8
0 (5) /ocfs2/test8r
1542 total
799 footprint
>>> 1578232 - 5120 - 107520 - 524288 - 101376 - 1024 - 21504
817400
jeff at jeff-laptop:./shared-du -E -k -c /ocfs2/*
0 /ocfs2/lost+found
88 /ocfs2/reflink
1024 (1024) /ocfs2/test1
1024 (1024) /ocfs2/test1r
21504 (21504) /ocfs2/test2
21504 (21504) /ocfs2/test2r
101376 (101376) /ocfs2/test3
101376 (101376) /ocfs2/test3r
40092 /ocfs2/test4
21508 (21508) /ocfs2/test4r
524288 (524288) /ocfs2/test5
524288 (524288) /ocfs2/test5r
107520 (107520) /ocfs2/test6
107520 (107520) /ocfs2/test6r
0 /ocfs2/test7
0 /ocfs2/test7r
5120 (5120) /ocfs2/test8
0 (5120) /ocfs2/test8r
1578232 total
817400 footprint
>>> 1621360142 - 5242880 - 110100480 - 536870912 - 103809024 - 22020096 - 1048576
842268174
jeff at jeff-laptop:./shared-du -E -b -c /ocfs2/*
3896 /ocfs2/bind_mount
3896 /ocfs2/lost+found
86414 /ocfs2/reflink
1048576 (1048576) /ocfs2/test1
1048576 (1048576) /ocfs2/test1r
22020096 (22020096) /ocfs2/test2
22020096 (22020096) /ocfs2/test2r
103809024 (103809024) /ocfs2/test3
103809024 (103809024) /ocfs2/test3r
41054208 /ocfs2/test4
22024192 (22024192) /ocfs2/test4r
536870912 (536870912) /ocfs2/test5
536870912 (536870912) /ocfs2/test5r
110100480 (110100480) /ocfs2/test6
110100480 (110100480) /ocfs2/test6r
1800 /ocfs2/test7 <<< inline file
1800 /ocfs2/test7r
5242880 (5242880) /ocfs2/test8
5242880 (5242880) /ocfs2/test8r
1621360142 total
842268174 footprint
Please review, any comments are appreciated!
Thanks,
-Jeff
^ permalink raw reply [flat|nested] 2+ messages in thread
* [Ocfs2-devel] du_enhancement v3
2010-03-08 15:30 [Ocfs2-devel] du_enhancement v3 jeff.liu
@ 2010-03-08 15:44 ` jeff.liu
0 siblings, 0 replies; 2+ messages in thread
From: jeff.liu @ 2010-03-08 15:44 UTC (permalink / raw)
To: ocfs2-devel
jeff.liu wrote:
> Hello,
>
> The following is the v3 patchset for du enhancement.
>
> The bugs fixed according to Tao's comments.
>
> For now, I faced with another problem need your guys help confirm:
> I skipped the inline file processing which its data mixed with metadata.
> IMHO, it has pretty minor effect for the precise measurement of footprint, as a result,
> the rebree process overhead is unworthy for inline file.
>
> Any thoughts or suggestions for this case?
>
>
> I run a few simple testing, it works:
>
>>>> 1542 - 1 - 21 - 99 - 512 - 105 - 5
> 799
> jeff at jeff-laptop:/shared-du -E -m -c /ocfs2/*
> 0 /ocfs2/lost+found
> 1 /ocfs2/reflink
> 1 (1) /ocfs2/test1
> 1 (1) /ocfs2/test1r
> 21 (21) /ocfs2/test2
> 21 (21) /ocfs2/test2r
> 99 (99) /ocfs2/test3
> 99 (99) /ocfs2/test3r
> 40 /ocfs2/test4
> 22 (22) /ocfs2/test4r
> 512 (512) /ocfs2/test5
> 512 (512) /ocfs2/test5r
> 105 (105) /ocfs2/test6
> 105 (105) /ocfs2/test6r
> 0 /ocfs2/test7
> 0 /ocfs2/test7r
> 5 (5) /ocfs2/test8
> 0 (5) /ocfs2/test8r
> 1542 total
> 799 footprint
>
>>>> 1578232 - 5120 - 107520 - 524288 - 101376 - 1024 - 21504
> 817400
> jeff at jeff-laptop:./shared-du -E -k -c /ocfs2/*
> 0 /ocfs2/lost+found
> 88 /ocfs2/reflink
> 1024 (1024) /ocfs2/test1
> 1024 (1024) /ocfs2/test1r
> 21504 (21504) /ocfs2/test2
> 21504 (21504) /ocfs2/test2r
> 101376 (101376) /ocfs2/test3
> 101376 (101376) /ocfs2/test3r
> 40092 /ocfs2/test4
> 21508 (21508) /ocfs2/test4r
> 524288 (524288) /ocfs2/test5
> 524288 (524288) /ocfs2/test5r
> 107520 (107520) /ocfs2/test6
> 107520 (107520) /ocfs2/test6r
> 0 /ocfs2/test7
> 0 /ocfs2/test7r
> 5120 (5120) /ocfs2/test8
> 0 (5120) /ocfs2/test8r
> 1578232 total
> 817400 footprint
>
>>>> 1621360142 - 5242880 - 110100480 - 536870912 - 103809024 - 22020096 - 1048576
> 842268174
> jeff at jeff-laptop:./shared-du -E -b -c /ocfs2/*
> 3896 /ocfs2/bind_mount
> 3896 /ocfs2/lost+found
> 86414 /ocfs2/reflink
> 1048576 (1048576) /ocfs2/test1
> 1048576 (1048576) /ocfs2/test1r
> 22020096 (22020096) /ocfs2/test2
> 22020096 (22020096) /ocfs2/test2r
> 103809024 (103809024) /ocfs2/test3
> 103809024 (103809024) /ocfs2/test3r
> 41054208 /ocfs2/test4
> 22024192 (22024192) /ocfs2/test4r
> 536870912 (536870912) /ocfs2/test5
> 536870912 (536870912) /ocfs2/test5r
> 110100480 (110100480) /ocfs2/test6
> 110100480 (110100480) /ocfs2/test6r
> 1800 /ocfs2/test7 <<< inline file
> 1800 /ocfs2/test7r
for the inline file, maybe a better way is show its shared size as well, it is straightforward, but
bypass split_extent() operation?
> 5242880 (5242880) /ocfs2/test8
> 5242880 (5242880) /ocfs2/test8r
> 1621360142 total
> 842268174 footprint
>
>
>
> Please review, any comments are appreciated!
>
> Thanks,
> -Jeff
>
> _______________________________________________
> Ocfs2-devel mailing list
> Ocfs2-devel at oss.oracle.com
> http://oss.oracle.com/mailman/listinfo/ocfs2-devel
^ permalink raw reply [flat|nested] 2+ messages in thread
end of thread, other threads:[~2010-03-08 15:44 UTC | newest]
Thread overview: 2+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2010-03-08 15:30 [Ocfs2-devel] du_enhancement v3 jeff.liu
2010-03-08 15:44 ` jeff.liu
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.