RE: Linux RAC and asynch io/IO slaves

  • From: "Parker, Matthew" <matthewp@xxxxxxxxxx>
  • To: <RROGERS@xxxxxxxxxxxxx>, <thump@xxxxxxxxxxxxxxxx>
  • Date: Fri, 14 Jan 2005 18:09:13 -0800


-----Original Message-----
From: oracle-l-bounce@xxxxxxxxxxxxx =
[mailto:oracle-l-bounce@xxxxxxxxxxxxx] On Behalf Of Ron Rogers
Sent: Friday, January 14, 2005 11:04 AM
To: thump@xxxxxxxxxxxxxxxx
Cc: cmarquez@xxxxxxxx; Parker, Matthew; oracle-l@xxxxxxxxxxxxx
Subject: RE: Linux RAC and asynch io/IO slaves

 I found this white paper from metalink to be helpful.

Asynchronous I/O (aio) on RedHat Advanced Server=20
Version 2.1 - 3.0    (Note:225751.1)


>>> "David" <thump@xxxxxxxxxxxxxxxx> 01/14/2005 1:05:03 PM >>>
Everything I have read advocates aio on RHAS 3.0.

I have heard also it can cause crashes and corruption, but found no meat =
to that statement no open bugs for 9205/9206 on rhas 3.0 Anyone with any =
hard facts as to real issues or successes?
Anyone done an IO slave implementation in Linux/RAC?
IO slaves in many ways are more robust and multi purpose than asych IO =
functionality...but I'm wondering how that performs with RAC.

Yes, I would and will test either or both before consideration with =
prod, but looking for others insights if they have been down the road.

>> Matthew,
>>>>bugs related to 9i RAC on Linux and asynch io
>> Can you point to any (bugs) specifically or should I just search=20
>> Metalink?
>> I was considering  enabling this for my 9iRAC/RHEL3ES...Oracle=20
>> (marketing, presentations) imply that I should be doing this.
>> Frankly Oracle Corp. takes much credit for helping RH/Linux get to=20
>> async-IO?
>> Thanks,
>> Chris Marquez
>> Oracle DBA
>> HEYMONitor(tm) -
>> "Oracle Monitoring & Alerting Solution"


Other related posts: