What is the value of your "filesystemio_options" parameter? Most likely, it isn't set to SETALL. If you were trying to have many simultaneous users, it would be a pain. On 02/26/2004 01:37:41 PM, Jared.Still@xxxxxxxxxxx wrote: > I've have 8i,9i and now 10g on RH 7.2 using all ext3 filesystems. > > This box has been in service for 2+ years. The only problems > with ext3 have been compatibility with Veritas NetBackup. > > Jared > > > > > > > Mladen Gogala <mgogala@xxxxxxxxxxxx> > Sent by: oracle-l-bounce@xxxxxxxxxxxxx > 02/25/2004 08:07 PM > Please respond to oracle-l > > > To: oracle-l@xxxxxxxxxxxxx > cc: > Subject: Re: Moving db to linux > > > Oracle supports Ext2 and OCFS, not Ext3. Ext3 has occasional problems > with corruption. As for OCFS, I'll give it a second look it on Saturday, > but it didn't look like it is capable of working without cluster. I should > be capable of linking it into my 2.4.25 kernel. As for async I/O, that is > not > a problem. I can always start a bunch of database writers instead of just > one. The problem with direct I/O is that without it, Linux is going to > put each and every block of the database in your buffer cache expanding > it, as necessary to accommodate such big demand. As PC's with infinite RAM > are pretty hard to find, that means that the RAM on your PC will be finite > in quantity and that Linux kernel will start throwing out other things > like > samba, network daemons, cupsd, nautilus and gdm to accommodate the > increased > demand for networking. Then someone approaches the keyboard and presses > enter or NE key. Linux will page nautilus, gdm, panels and related stuff > right back in, shrinking the buffer pool. Then oracle tries to do > transaction, Linux expands buffer pool....We've got what's known as > "circulus viciosus" or "vicious circle". To the buyers of the IBM > equipment from few decades ago, it is also known as "the spiral of > death". The funny thing is that all that buffering is completely > unnecessary because oracle does its own buffering in SGA, so the Linux > buffering is redundant and inefficient. It does improve BCHR, though, > which is of paramount importance, as is well known on this group. > The only way to avoid costly, and unnecessary double buffering is to make > Linux read and write database files bypassing the buffer cache. That is > the very definition of direct I/O. Now, we're back to square 2. > > > On 02/25/2004 10:22:09 PM, zhu chao wrote: > > Isn't ext3 good enough? It provide JFS recovery performance and AIO/DIO > on > > redhat linux.(On redhat 3, aio seems still have problems, but seems > mature > > on 2.1) > > > > Oracle metalink only support ext3/ocfs etc. Seems no jfs support. > > > > Regards > > Zhu Chao. > > > > ----- Original Message ----- > > From: "Mladen Gogala" <mladen@xxxxxxxxxxxxxxx> > > To: <oracle-l@xxxxxxxxxxxxx> > > Sent: Thursday, February 26, 2004 3:12 AM > > Subject: Re: Moving db to linux > > > > > > > It should already be in the distribution. Type "rpm -qa|grep jfsutils" > and > > > you should get an answer like this: > > > Yogi> rpm -qa|grep jfs > > > jfsutils-1.0.17-6 > > > Yogi> > > > > > > JFS is a standard part of the kernel. The "utils" package contains all > > > you need for for checking it, making it and alike. Here is contents of > the > > > standard distribution on RH-9: > > > > > > Yogi> rpm -ql jfsutils-1.0.17-6 > > > /sbin/defragfs > > > /sbin/extendfs > > > /sbin/fsck.jfs > > > /sbin/logdump > > > /sbin/logredo > > > /sbin/mkfs.jfs > > > /sbin/xchkdmp > > > /sbin/xchklog > > > /sbin/xpeek > > > /usr/share/doc/jfsutils-1.0.17 > > > /usr/share/doc/jfsutils-1.0.17/AUTHORS > > > /usr/share/doc/jfsutils-1.0.17/COPYING > > > /usr/share/doc/jfsutils-1.0.17/ChangeLog > > > /usr/share/doc/jfsutils-1.0.17/INSTALL > > > /usr/share/doc/jfsutils-1.0.17/NEWS > > > /usr/share/doc/jfsutils-1.0.17/README > > > /usr/share/man/man8/fsck.jfs.8.gz > > > /usr/share/man/man8/logdump.8.gz > > > /usr/share/man/man8/logredo.8.gz > > > /usr/share/man/man8/mkfs.jfs.8.gz > > > /usr/share/man/man8/xchkdmp.8.gz > > > /usr/share/man/man8/xchklog.8.gz > > > /usr/share/man/man8/xpeek.8.gz > > > Yogi> > > > > > > Can you please check that you have it? I'm very eager to learn, > because > > that's > > > precisely the problem I reported to oracle with 10g. The result is > bug# > > 3446411 > > > (public) and I'm in the process of disembowelling my existing RH-9 box > and > > > turning it into ES 3.0, because oracle told me that the ES 3.0 kernel > may > > contain > > > the necessary ingredients to make it work. > > > > > > > > > On 02/25/2004 01:54:50 PM, "Nelson, Allan" wrote: > > > > Really, do you know where I can get JFS for RH AS 3.0? I can work > from > > > > tarballs as well. > > > > > > > > Thanks > > > > > > > > -----Original Message----- > > > > From: oracle-l-bounce@xxxxxxxxxxxxx > > > > [mailto:oracle-l-bounce@xxxxxxxxxxxxx] On Behalf Of Mladen Gogala > > > > Sent: Wednesday, February 25, 2004 12:05 PM > > > > To: oracle-l@xxxxxxxxxxxxx > > > > Subject: Re: Moving db to linux > > > > > > > > > > > > My advice is to use JFS for database files because on 9.2.0.4, you > can > > > > use direct I/O, which is supported by JFS. As far as the big files > are > > > > concerned, JFS supports them without any problem. > > > > > > > > On 02/25/2004 12:22:58 PM, "Nelson, Allan" wrote: > > > > > Im, going to be moving a 220 GB 8.1.7.4 database from HP-UX to Red > Hat > > > > > AS 3.0 running 9.2.0.4.0. I'm told that we have to use > export/import > > > > to > > > > > get that job done. I plan to work through pipes and use split to > keep > > > > > the file sizes to 2 GB or less to aviod large file issues between > the > > > > > platforms. Does anyone know of other issues that may trip me up? > > > > > > > > > > Thanks In Advance > > > > > > > > > > > Allan L. Nelson > > > > > > Oracle DBA > > > > > > M-I L.L.C. > > > > > > (832) 295-2238 office > > > > > > (832) 351-4180 fax > > > > > > anelson@xxxxxxxx <mailto:anelson@xxxxxxxx> > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > ________________________________________________________________________ > > > > ______ > > > > > This email is intended solely for the person or entity to which it > is > > > > addressed and may contain confidential and/or privileged > information. > > > > Copying, forwarding or distributing this message by persons or > entities > > > > other than the addressee is prohibited. If > > > > > > > > > > > > > > ---------------------------------------------------------------- > > > > Please see the official ORACLE-L FAQ: http://www.orafaq.com > > > > ---------------------------------------------------------------- > > > > To unsubscribe send email to: oracle-l-request@xxxxxxxxxxxxx > > > > put 'unsubscribe' in the subject line. > > > > -- > > > > Archives are at //www.freelists.org/archives/oracle-l/ > > > > FAQ is at //www.freelists.org/help/fom-serve/cache/1.html > > > > ----------------------------------------------------------------- > > > > > > > > > > > > > > > ____________________________________________________________________________ > > __ > > > > This email is intended solely for the person or entity to which it > is > > addressed and may contain confidential and/or privileged information. > > Copying, forwarding or distributing this message by persons or entities > > other than the addressee is prohibited. If you have received this email > in > > error, please contact the sender immediately and delete the material > from > > any computer. This email may have been monitored for policy compliance. > > [021216] > > > > > > > > ---------------------------------------------------------------- > > > > Please see the official ORACLE-L FAQ: http://www.orafaq.com > > > > ---------------------------------------------------------------- > > > > To unsubscribe send email to: oracle-l-request@xxxxxxxxxxxxx > > > > put 'unsubscribe' in the subject line. > > > > -- > > > > Archives are at //www.freelists.org/archives/oracle-l/ > > > > FAQ is at //www.freelists.org/help/fom-serve/cache/1.html > > > > ----------------------------------------------------------------- > > > > > > > ---------------------------------------------------------------- > > > Please see the official ORACLE-L FAQ: http://www.orafaq.com > > > ---------------------------------------------------------------- > > > To unsubscribe send email to: oracle-l-request@xxxxxxxxxxxxx > > > put 'unsubscribe' in the subject line. > > > -- > > > Archives are at //www.freelists.org/archives/oracle-l/ > > > FAQ is at //www.freelists.org/help/fom-serve/cache/1.html > > > ----------------------------------------------------------------- > > > > > > > > > > ---------------------------------------------------------------- > > Please see the official ORACLE-L FAQ: http://www.orafaq.com > > ---------------------------------------------------------------- > > To unsubscribe send email to: oracle-l-request@xxxxxxxxxxxxx > > put 'unsubscribe' in the subject line. > > -- > > Archives are at //www.freelists.org/archives/oracle-l/ > > FAQ is at //www.freelists.org/help/fom-serve/cache/1.html > > ----------------------------------------------------------------- > > > > -- > Mladen Gogala > Oracle DBA > ---------------------------------------------------------------- > Please see the official ORACLE-L FAQ: http://www.orafaq.com > ---------------------------------------------------------------- > To unsubscribe send email to: oracle-l-request@xxxxxxxxxxxxx > put 'unsubscribe' in the subject line. > -- > Archives are at //www.freelists.org/archives/oracle-l/ > FAQ is at //www.freelists.org/help/fom-serve/cache/1.html > ----------------------------------------------------------------- > > > ---------------------------------------------------------------- Please see the official ORACLE-L FAQ: http://www.orafaq.com ---------------------------------------------------------------- To unsubscribe send email to: oracle-l-request@xxxxxxxxxxxxx put 'unsubscribe' in the subject line. -- Archives are at //www.freelists.org/archives/oracle-l/ FAQ is at //www.freelists.org/help/fom-serve/cache/1.html -----------------------------------------------------------------