Oracle FAQ Your Portal to the Oracle Knowledge Grid

Home -> Community -> Mailing Lists -> Oracle-L -> Re: Oracle (11g) on Solaris X86 - bad move?

Re: Oracle (11g) on Solaris X86 - bad move?

From: Mark Brinsmead <>
Date: Fri, 14 Dec 2007 21:54:40 -0700
Message-ID: <>


   I have just done some research on this subject for a customer.

   Let me first note that you may have blended two platforms into one. Oracle supports "Solaris-x86" and "Solaris-x86_64" as two distinct platforms -- 32-bit and 64-bit respectively.

   The "Solaris-x86" (32-bit) platform is dead for all practical purposes. In the Metalink "Certify" matrix, there is a clear statement that 10gR2 is the terminal release for "Solaris-x86". Further, there is no plan to *ever*release the patchset for 32-bit solaris, and Oracle is not saying when (or whether) they plan to release for same.

   The 64-bit platform is at least a little different. *has* been released, as I recall, for 64-bit Solaris-x86 ("Solaris-x86_64"), and is planned for release some time in 2008. There has been no statement made regarding when (or whether!) 11g will be released for this platform.

   My client is prepared to gamble that 11g *will* be released for Solaris-x86_64, eventually, at least. (They have no interest in 11g for at least a year anyway.) Of course, I have advised caution -- personally I would not place a large wager on the question of whether the platform survives to see the release of 11g, but it might.

On Dec 10, 2007 1:23 AM, John Hallas <> wrote:

> Thanks Alex,
> Yours was the only response to the list but I had 3 private messages all
> basically agreeing with what I see as a major concern.
> One sentence seems to summarise the situation "However, the lack of
> direction and support from Oracle on certifying EBS on Solaris-x86 has
> been a barrier for us to come to the final decision".
> Whilst not wanting to go too specifically into our situation the client
> has selected Oracle as the database of choice and the x86 chip as best
> bang for the buck. I personally think that the o/s should be Linux, but
> that requires a major sea-change for the client. Before I raised the
> issue I wanted to be sure that I was not a voice in the wilderness and
> easily shot down. I don't think that is the case.
> The application code is pretty o/s independent anyway as you pointed out
> it should be.
> John
> ...

-- Mark Brinsmead
  Senior DBA,
  The Pythian Group

Received on Fri Dec 14 2007 - 22:54:40 CST

Original text of this message