Instance Crashed !!

Hi, We are using IDS FC6 on Sun Solaris. We've allocated 8GB of Memory (SHMTOTAL) to one of our instance. When its memory approached 5.2 GB, it remained stable for around 1.5 hrs. Afterwards, the instance stopped response (even we could not access it using dbaccess). The checkpoint also took time, and at the time of instance (hang), the checkpoint could not be genrated. Is there any OS related parameter that can help/avoid/overlap the resources for Instance related configuration ? Is there any max limit or a particular ration of SHMTOTAL to currently used memory for any instance ? Thanks.

Did the instance actually crash? Is there an af file generated or an assertion failure message in the online.log? FC6 is a sub release of the engine... Is it version 10, 11.1, 11.5, 11.7 or something older? I assume you have enough physical RAM. You could use ipcs to have a look at what other processes are using shared memory.... Jarrod -----Original Message----- From: ids-bounces@iiug.org [mailto:ids-bounces@iiug.org] On Behalf Of SHAHZAD SALAM KASI Sent: Friday, 21 January 2011 10:19 p.m. To: ids@iiug.org Subject: Instance Crashed !! [22510] Hi, We are using IDS FC6 on Sun Solaris. We've allocated 8GB of Memory (SHMTOTAL) to one of our instance. When its memory approached 5.2 GB, it remained stable for around 1.5 hrs. Afterwards, the instance stopped response (even we could not access it using dbaccess). The checkpoint also took time, and at the time of instance (hang), the checkpoint could not be genrated. Is there any OS related parameter that can help/avoid/overlap the resources for Instance related configuration ? Is there any max limit or a particular ration of SHMTOTAL to currently used memory for any instance ? Thanks. ******************************************************************************* Forum Note: Use "Reply" to post a response in the discussion forum. DISCLAIMER This email contains information that is confidential and which may be legally privileged. If you have received this email in error, please notify the sender immediately and delete the email. This email is intended solely for the use of the intended recipient and you may not use or disclose this email in any way.

We are using 11.50.FC6.

Hi, We re-started the instance, as it was'nt responsive. There was'nt any error message in log file as well. There was more than 3 GB of Physical Memory available at that time. Thanks for the response.

The memory limit for .FC (64BIT) releases is 4TB. How much physical memory does the system have? How much swap space? Also, what is the full version? Art On Jan 21, 2011 2:19 AM, "SHAHZAD SALAM KASI" <skasi@i2cinc.com> wrote: > Hi, > We are using IDS FC6 on Sun Solaris. > We've allocated 8GB of Memory (SHMTOTAL) to one of our instance. When its > memory approached 5.2 GB, it remained stable for around 1.5 hrs. Afterwards, > the instance stopped response (even we could not access it using dbaccess). > The checkpoint also took time, and at the time of instance (hang), the > checkpoint could not be genrated. > Is there any OS related parameter that can help/avoid/overlap the resources > for Instance related configuration ? > > Is there any max limit or a particular ration of SHMTOTAL to currently used > memory for any instance ? > > Thanks. > > > ******************************************************************************* > Forum Note: Use "Reply" to post a response in the discussion forum. > --001636c5b45f9bea0b049a5eafbb

Physical Memory of machine is 32 GB. Swap space is 24 GB. Using IDS 11.50.FC6 Also, is it better to allocate more memory (initially) or in chunks (for OLTP Application) ? Thanks

Best to allocate as much memory initially as you are likely to use day to day. See my other responses about the crash. Art On Jan 21, 2011 10:43 AM, "SHAHZAD SALAM KASI" <skasi@i2cinc.com> wrote: > Physical Memory of machine is 32 GB. > Swap space is 24 GB. > > Using IDS 11.50.FC6 > > Also, is it better to allocate more memory (initially) or in chunks (for OLTP > Application) ? > > Thanks > > > ******************************************************************************* > Forum Note: Use "Reply" to post a response in the discussion forum. > --485b393aafa511ff93049a5eecdd

What must be the ratio of Resident Memory & Shared Memory that we can allocate initially ? What is the impact of # of Buffers & its size to the size of Resident Memory. Thanks.

The resident segment is mostly the buffer cache, so that determines its size. The virtual segments contain the smaller caches like the dictionary cache etc. And working memory for sessions to use. There is no ratio between the two. Art On Jan 21, 2011 11:12 AM, "SHAHZAD SALAM KASI" <skasi@i2cinc.com> wrote: > What must be the ratio of Resident Memory & Shared Memory that we can allocate > initially ? What is the impact of # of Buffers & its size to the size of > Resident Memory. > > Thanks. > > > ******************************************************************************* > Forum Note: Use "Reply" to post a response in the discussion forum. > --20cf3054a5b1ae7d6e049a61267d

On 21/01/2011 09:19, SHAHZAD SALAM KASI wrote: > Hi, > We are using IDS FC6 on Sun Solaris. > We've allocated 8GB of Memory (SHMTOTAL) to one of our instance. When its > memory approached 5.2 GB, it remained stable for around 1.5 hrs. Afterwards, > the instance stopped response (even we could not access it using dbaccess). > The checkpoint also took time, and at the time of instance (hang), the > checkpoint could not be genrated. > Is there any OS related parameter that can help/avoid/overlap the resources > for Instance related configuration ? > > Is there any max limit or a particular ration of SHMTOTAL to currently used > memory for any instance ? Sounds to me more like your logical logs are full :o) -- Cheers, Obnoxio The Clown http://obotheclown.blogspot.com I will now proceed to pleasure myself with this fish.

I agree with Obnoxio. looks like you can't write any transactions anymore because logs are full.