| Oracle FAQ | Your Portal to the Oracle Knowledge Grid | |
 |
 | |||
| Oracle FAQ | Your Portal to the Oracle Knowledge Grid | |
|
| |||
Home -> Community -> Usenet -> c.d.o.server -> Re: Configuring Veritas Quick I/O & Oracle 8
>From the UNIX side of it, there does not seem to be any particular re-tuning
>to be done when using Veritas file systems & Quick I/O (neither when using
>raw devices if we come to that). Hence I am suspecting that it is either the
>type of test being done which is accentuating the benefits of normal file
>systems over direct I/O to files, or that there is some Oracle tuneables
>that have not been properly set.
>I would really appreciate any help or advice regarding the matter.
>
I'm a sys admin at a large SAP/Oracle shop. We currently are running vxfs filesystem. We too are exploring Quick I/O. For your setup my guess is you have a large amount of RAM. Normal UFS under Solaris will use basically all available memory as a file system buffer. So what happens on your system is a lot of the database is getting cached in the filesystem buffer. This is great if you have enough memory. But if you don't have enough or your database is very large and heavily loaded (like ours) you will soon see the system spending more and more time trying to free up pages of memory to handle all the i/o. At this point the system starts to bog down, spending more time paging things out then actually doing any real work. With the proper mount options, (forcedirectio for UFS, mincache=direct, convosync=direct for vxfs) you can bypass the buffering for filesystem. That is what we had to do. QIO also bypasses the file system buffers. The latest version of QIO does support cacheing. Have had a chance to try it but it does look very cool. QIO also gives you support for kernel level async i/o which you don't have when just using file systems. It would be interesting to see how long your test would run if you mounted your ufs file systems up with the forcedirectio option.
HTH,
john
 |
 |