Re: Raw Disk Partition vs. Unix File System (Oracle 6)
Date: Tue, 17 Aug 93 17:18:04 GMT
Message-ID: <1993Aug17.171804.15624_at_exlog.com>
In article <1993Aug16.224048.15479_at_lmpsbbs.comm.mot.com> cds016_at_isadmin1.comm.mot.com (David Schmitt) writes:
>In article <CBuyrL.DLo_at_cbnewse.cb.att.com>,
>richard.dib <cafe_at_cbnewse.cb.att.com> wrote:
>>
>>Here are my questions:
>>
>>- Does anybody have any experience with Oracle using a raw disk partitition
>> instead of the Unix file system? Could you share your experiences?
>
>I've used raw devices with Oracle 6 for several years, and only had one
>bad experience (when a system admin put swap space on top of a data file).
>It is more difficult to administer in many environments, since it is hard
>to adjust disk layout on the fly (you typically have to reslice your disk).
>However, with planning I haven't found this to be that much of a problem.
Planning is the key word, My major complaint about people that stand up and say "Lets Go Raw" is that they often are not willing to pay the price to implement correctly.
OK, How are YOU going to back up your database?
"Oh we thought you would do that"
Oh and don't forget to shutdown the database.
I never heard from them again.
>>- What are the benefits? Is the improvement in performance worth the work?
>
>Benefits are improved integrity (don't have to worry about whether UNIX
>buffers were flushed at the time of a crash), and performance improvements.
>I have never benchedmarked the difference in performance, but it makes
>sense to me to believe there is one.
I'm not sure I agree with the integrity issue. If Oracle uses the O_SYNC options on their writes (they SAY they do) then you should be ensured that data got to atleast the LOGFILE intact and thats all you really need.
I'm also not sure I agree with the performance improvment statement in ALL cases. Seems to me that if you are dealing with large FS blocksizes ie) 32K-64K, then your improvement should be if not negligable then atleast greatly reduces. Further if your dealing with SunOS and have a large amount of RAM that is used as a massive transparant buff cache then the FS could be a win. Depenting on the setup up they may beable to get near raw performance on a cooked db.
[...]
-- Regards, Lee E. Parsons Baker Hughes Inteq, Inc Oracle Database Administrator lparsons_at_exlog.comReceived on Tue Aug 17 1993 - 19:18:04 CEST