From mboxrd@z Thu Jan 1 00:00:00 1970 From: Mark Fasheh Date: Mon Aug 2 15:11:51 2004 Subject: [Ocfs2-devel] [Patch] how about remove the sync in ocfs_file_open? In-Reply-To: <3ACA40606221794F80A5670F0AF15F8404FFE3CD@pdsmsx403> References: <3ACA40606221794F80A5670F0AF15F8404FFE3CD@pdsmsx403> Message-ID: <20040802200334.GH8008@ca-server1.us.oracle.com> List-Id: MIME-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit To: ocfs2-devel@oss.oracle.com On Thu, Jul 29, 2004 at 02:19:57AM +0800, Ling, Xiaofeng wrote: > >From the data below, this sync cause the performance of writting from > second time downgrade(13M compare to 544M for 128K file). > and the sync has no use here. Well, since it's really easy to revert, we're going to try this one and see if it breaks anything multi-node. If it does, you can be sure we'll put the sync_inode back :) --Mark > Index: file.c > =================================================================== > --- file.c (revision 1308) > +++ file.c (working copy) > @@ -300,7 +301,6 @@ > have_oin_sem = 0; > } > > - status = ocfs_sync_inode(inode); > > leave: > if (have_oin_sem) { > > > before remove > ----------------------------------------------------------------- > ./ben -0 -p10 -a > create file /ocfs/ocfs_test25032 of size 128K successfully > total:128.000K Time:0.562 Rate:227758KB > total:128.000K Time:9.795 Rate:13068KB > total:128.000K Time:9.687 Rate:13214KB > total:128.000K Time:8.999 Rate:14224KB > total:128.000K Time:9.817 Rate:13039KB > total:128.000K Time:9.802 Rate:13059KB > total:128.000K Time:9.146 Rate:13995KB > total:128.000K Time:9.646 Rate:13270KB > total:128.000K Time:9.655 Rate:13257KB > total:128.000K Time:9.028 Rate:14178KB > total:1280.000K Time:87.329 Rate:14657KB > > after remove > ------------------------------------------------------------------------ > ------------------ > ./ben -0 -p10 -a > create file /ocfs/ocfs_test24613 of size 128K successfully > total:128.000K Time:0.292 Rate:438356KB > total:128.000K Time:0.235 Rate:544681KB > total:128.000K Time:0.209 Rate:612440KB > total:128.000K Time:0.225 Rate:568889KB > total:128.000K Time:0.207 Rate:618357KB > total:128.000K Time:0.219 Rate:584475KB > total:128.000K Time:0.205 Rate:624390KB > total:128.000K Time:0.207 Rate:618357KB > total:128.000K Time:0.204 Rate:627451KB > total:128.000K Time:0.204 Rate:627451KB > total:1280.000K Time:3.083 Rate:415180KB > ------------------- > > > Index: file.c > =================================================================== > --- file.c (revision 1308) > +++ file.c (working copy) > @@ -300,7 +301,6 @@ > have_oin_sem = 0; > } > > - status = ocfs_sync_inode(inode); > > leave: > if (have_oin_sem) { > > > > > Intel China Software Lab. > iNet: 8-752-1243 > 021-52574545-1243(O) > xfling@users.sourceforge.net > Opinions are my own and don't represent those of my employer > _______________________________________________ > Ocfs2-devel mailing list > Ocfs2-devel@oss.oracle.com > http://oss.oracle.com/mailman/listinfo/ocfs2-devel -- Mark Fasheh Software Developer, Oracle Corp mark.fasheh@oracle.com