Re: Is database read activity primary sequential or random?
From: Mladen Gogala <mgogala_at_yahoo.com>
Date: Fri, 17 Apr 2015 13:26:26 -0400
Message-ID: <55314242.2070108_at_yahoo.com>
On 04/17/2015 01:04 AM, MARK BRINSMEAD wrote:
> But that doesn't mean that the database isn't *trying* to do
> sequential access. Nor does it mean it doesn't *think* that it
> actually is doing sequential access.
>
> There are a few more cases like this. Many SAN-attached disk arrays
> these days are highly virtualized, and offer "cool" services like thin
> provisioning and snapshot capabilities. When the disk array does
> relocate-on-write (a common way to support snapshots) it may not take
> a long time before IO operations that _/should/_ be sequential have
> become almost entirely random. Again, the database won't know this --
> but the disk heads certainly will.
Agreed. But I don't really care what does database think it's doing, I only care about what the database is actually doing. And the answer is that you can never tell.
Date: Fri, 17 Apr 2015 13:26:26 -0400
Message-ID: <55314242.2070108_at_yahoo.com>
On 04/17/2015 01:04 AM, MARK BRINSMEAD wrote:
> But that doesn't mean that the database isn't *trying* to do
> sequential access. Nor does it mean it doesn't *think* that it
> actually is doing sequential access.
>
> There are a few more cases like this. Many SAN-attached disk arrays
> these days are highly virtualized, and offer "cool" services like thin
> provisioning and snapshot capabilities. When the disk array does
> relocate-on-write (a common way to support snapshots) it may not take
> a long time before IO operations that _/should/_ be sequential have
> become almost entirely random. Again, the database won't know this --
> but the disk heads certainly will.
Agreed. But I don't really care what does database think it's doing, I only care about what the database is actually doing. And the answer is that you can never tell.
-- Mladen Gogala Oracle DBA http://mgogala.freehostia.com -- http://www.freelists.org/webpage/oracle-lReceived on Fri Apr 17 2015 - 19:26:26 CEST