Re: Oracle DB Best Practices on Linux x86_64
- From: Justin Mungal <justin@xxxxxxx>
- To: Timur Akhmadeev <timur.akhmadeev@xxxxxxxxx>
- Date: Tue, 29 Jul 2014 03:28:00 -0500
On systems with low-load, I just use AMM and leave HugePages off.
On Tue, Jul 29, 2014 at 3:08 AM, Timur Akhmadeev <timur.akhmadeev@xxxxxxxxx>
wrote:
> Well I think it's the only thing that is a must on Linux box these days
>
>
> On Tuesday, July 29, 2014, Justin Mungal <justin@xxxxxxx> wrote:
>
>> Yes but HugePages is one of those "use it if you actually need it" types
>> of things... IMHO...
>>
>>
>> On Tue, Jul 29, 2014 at 2:57 AM, Timur Akhmadeev <
>> timur.akhmadeev@xxxxxxxxx> wrote:
>>
>>> >Use the automation features of the database like Automatic Memory
>>> Management (AMM)
>>>
>>> Funny you mention ORAchk and AMM - the first will specifically recommend
>>> huge pages which is not possible to configure with AMM in place
>>>
>>> On Monday, July 28, 2014, Seth Miller <sethmiller.sm@xxxxxxxxx> wrote:
>>>
>>>> Justin,
>>>>
>>>> I would suggest running the latest ORAchk (1268927.2) on all of the
>>>> systems in question. The results of the tool will show violations of and
>>>> recommendations for best practices. This will also address some of the
>>>> items you listed, like multiplexing control files and redo logs. MOS
>>>> recommends running this monthly.
>>>>
>>>> Enterprise Manager can be used to monitor things like patch level and
>>>> make opening support tickets MUCH easier and in some cases automatic.
>>>>
>>>> Use the automation features of the database like Automatic Memory
>>>> Management (AMM) which implicitly uses Automatic Shared Memory Management
>>>> (ASMM), Automatic Undo Managment, Automatic Segment Space Management
>>>> (ASSM), Automatic Diagnostic Repository (ADR) and the maintenance jobs to
>>>> name a few.
>>>>
>>>> Anything left over after these things is probably based on the database
>>>> requirements and needs to be defined by the customer.
>>>>
>>>> Seth Miller
>>>>
>>>>
>>>>
>>>> On Mon, Jul 28, 2014 at 3:16 AM, Justin Mungal <justin@xxxxxxx> wrote:
>>>>
>>>>> I have an interesting request from a customer to review the Oracle
>>>>> configuration on a couple servers for best practices. I've been checking
>>>>> things over; I haven't had a chance to talk to the customer yet, but
>>>>> seeing
>>>>> from the load I'm guessing they aren't in production yet.
>>>>>
>>>>> Right now I've just looked at the very basic things that tend to come
>>>>> back and bite later on, or are just good things to configure in general:
>>>>> backups, multiplexed control/redo logs, memory settings, patch levels, log
>>>>> rotation, etc... but then again not all of them may be necessary,
>>>>> depending
>>>>> on the customer's needs.
>>>>>
>>>>> Has anyone created some kind of generic best practices document, or a
>>>>> checklist of some kind? Arup Nanda's document from 2007 has some good
>>>>> points, so I would be looking for something like that I suppose.
>>>>>
>>>>> Thanks,
>>>>> Justin
>>>>>
>>>>
>>>>
>>>
>>> --
>>> Regards
>>> Timur Akhmadeev
>>>
>>>
>>
>
> --
> Regards
> Timur Akhmadeev
>
>
Other related posts:
