Re: limit oracle memory
Date: Mon, 27 Jul 2009 13:13:06 GMT
Message-ID: <bu9r65doi4vgs82101c9e1t41s3rkc1dms_at_4ax.com>
On Sun, 26 Jul 2009 18:43:22 +0200, "Matthias Hoys" <anti_at_spam.com> wrote:
>
>"NetComrade" <netcomradeNSPAM_at_bookexchange.net> wrote in message
>news:cdso65dqr108ofso0upv9iin5gl88780e7_at_4ax.com...
>> On Sat, 25 Jul 2009 07:37:17 -0700 (PDT), Mark D Powell
>> <Mark.Powell_at_eds.com> wrote:
>>
>>>How much real memory does the server have?
>>
>>
>>>What memory settings have you used for each database instance?
>> 16G ram, 16G swap
>>
>>>Where do your applications run?
>> on other servers.
>>
>>>What type of applications are these databases supporting?
>> mostly jboss.
>>
>>>Any applications other than databases and normal OS processes running
>>>on the server?
>>
>> no, but there are tons of scripts/"batch jobs"
>>
>>>Have the administrator take a look at the sar data history. This
>>>should help determine if this was a one time issue caused by something
>>>running or if you have had a trend where spare capacity was declining.
>>
>> It was a one time issue. It happens once every few months and very
>> quickly, if it was a trend swap monitors would've caught it.
>>
>> SGA max is under 2G
>> PGA target is 2G
>> sessions are limited to 555, but normally hovers around 100.
>> total processes hover around 140-150
>>
>> -a
>You say "it happens once every few months". Maybe there is some script
>scheduled in cron that runs at that time and causes the problems? Or a job
>scheduled in Oracle?
We weren't able to identify such a cron.
.......
We run Oracle 9iR2,10gR2, 10g2RAC on RH4/RH5 and Solaris 10 (Sparc)
We use RMAN and remote catalog for backups
Received on Mon Jul 27 2009 - 08:13:06 CDT