Managing memory in shared mode

Dear all,
1. Is the oracle shared server mode assign equal memory each session?
And where i can find the document about managing memory of shared server mode ?
Thks for advance
Chara

Thanks hsawwan
I cannot find the documents
Note: 461063.1 - HR and Financials Applications Installed as Full Instead of Shared
Note: 461063.1 - Changing Oracle Human Resources Installation from FULL HR/FULL Payroll Install to SHARED HR/SHARED Payroll Install
in Metalink. Do you have any idea if thos documents still available?
Those documents change HRMS from Install to Shared Mode when the application is already installed.
Is any way to install HRMS in shared mode before install the application?
Regards
Alvaro

Similar Messages

System.InsufficientMemoryException: Failed to allocate a managed memory buffer of 268435456 bytes. The amount of available memory may be low. --- System.OutOfMemoryException: Exception of type 'System.OutOfMemoryException' was thrown.

Appfabric 1.1 server setup on 3 Windows server 2008 R2 machines.
Client windows 7 64 bit.
Because there is no bulkupdate ,we are trying to persist around 4000 objects wrapped in a CLR object.
[Serializable]
public class CacheableCollection<T> : ICacheable, IEnumerable<T>
where T : class, ICacheable
[DataMember]
private Dictionary<string, T> _items;
public IEnumerator<T> GetEnumerator()
return _items.Values.GetEnumerator();
IEnumerator IEnumerable.GetEnumerator()
return _items.Values.GetEnumerator();
[DataMember]
public string CacheKey { get; private set; }
public T this[string cacheKey] { get { return _items[cacheKey]; } }
public CacheableCollection(string cacheKey, T[] items)
if (string.IsNullOrWhiteSpace(cacheKey))
throw new ArgumentNullException("cacheKey", "Cache key not specified.");
if (items == null || items.Length == 0)
throw new ArgumentNullException("items", "Collection items not specified.");
this.CacheKey = cacheKey;
_items = items.ToDictionary(p => p.CacheKey, p => p);
We tried with the following options on server and client
Server:
<advancedProperties>
<partitionStoreConnectionSettings leadHostManagement="false" />
<securityProperties mode="None" protectionLevel="None">
<authorization>
<allow users="[email protected]" />
<allow users="[email protected]" />
</authorization>
</securityProperties>
<transportProperties maxBufferSize="500000000" />
</advancedProperties>
Client:
<transportProperties connectionBufferSize="131072" maxBufferPoolSize="500000000"
maxBufferSize="838860800" maxOutputDelay="2" channelInitializationTimeout="60000"
receiveTimeout="600000"/>
I see different people experiencing different memory size issues. What is the actual memory limit of an object that can be pushed to Appfabric.
Can some one please help ?
Stack trace:
Test method Anz.Cre.Pdc.Bootstrapper.Test.LoaderFuncCAOTest.AppFabPushAndRetrieveData threw exception:
System.InsufficientMemoryException: Failed to allocate a managed memory buffer of 268435456 bytes. The amount of available memory may be low. ---> System.OutOfMemoryException: Exception of type 'System.OutOfMemoryException' was thrown.
System.Runtime.Fx.AllocateByteArray(Int32 size)
System.Runtime.Fx.AllocateByteArray(Int32 size)
System.Runtime.InternalBufferManager.PooledBufferManager.TakeBuffer(Int32 bufferSize)
System.Runtime.BufferedOutputStream.ToArray(Int32& bufferSize)
System.ServiceModel.Channels.BufferedMessageWriter.WriteMessage(Message message, BufferManager bufferManager, Int32 initialOffset, Int32 maxSizeQuota)
System.ServiceModel.Channels.BinaryMessageEncoderFactory.BinaryMessageEncoder.WriteMessage(Message message, Int32 maxMessageSize, BufferManager bufferManager, Int32 messageOffset)
System.ServiceModel.Channels.FramingDuplexSessionChannel.EncodeMessage(Message message)
System.ServiceModel.Channels.FramingDuplexSessionChannel.OnSendCore(Message message, TimeSpan timeout)
System.ServiceModel.Channels.TransportDuplexSessionChannel.OnSend(Message message, TimeSpan timeout)
System.ServiceModel.Channels.OutputChannel.Send(Message message, TimeSpan timeout)
Microsoft.ApplicationServer.Caching.WcfClientChannel.SendMessage(EndpointID endpoint, Message message, TimeSpan timeout, WaitCallback callback, Object state, Boolean async)
Microsoft.ApplicationServer.Caching.WcfClientChannel.Send(EndpointID endpoint, Message message, TimeSpan timeout)
Microsoft.ApplicationServer.Caching.WcfClientChannel.Send(EndpointID endpoint, Message message)
Microsoft.ApplicationServer.Caching.DRM.SendRequest(EndpointID address, RequestBody request)
Microsoft.ApplicationServer.Caching.RequestBody.Send()
Microsoft.ApplicationServer.Caching.DRM.SendToDestination(RequestBody request, Boolean recordRequest)
Microsoft.ApplicationServer.Caching.DRM.ProcessRequest(RequestBody request, Boolean recordRequest)
Microsoft.ApplicationServer.Caching.DRM.ProcessRequest(RequestBody request, Object session)
Microsoft.ApplicationServer.Caching.RoutingClient.SendMsgAndWait(RequestBody reqMsg)
Microsoft.ApplicationServer.Caching.DataCache.SendReceive(RequestBody reqMsg)
Microsoft.ApplicationServer.Caching.DataCache.ExecuteAPI(RequestBody reqMsg)
Microsoft.ApplicationServer.Caching.DataCache.InternalPut(String key, Object value, DataCacheItemVersion oldVersion, TimeSpan timeout, DataCacheTag[] tags, String region)
Microsoft.ApplicationServer.Caching.DataCache.Put(String key, Object value, String region)
Anz.Cre.Pdc.DataCache.DataCacheAccess.Put[T](String cacheName, String regionName, T value) in C:\SVN\2.3_Drop3\app\Src\Anz.Cre.Pdc.DataCache\DataCacheAccess.cs: line 141
Anz.Cre.Pdc.DataCache.DataCacheAccess.Put[T](String cacheName, String regionName, Boolean flushRegion, T value) in C:\SVN\2.3_Drop3\app\Src\Anz.Cre.Pdc.DataCache\DataCacheAccess.cs: line 372
Anz.Cre.Pdc.Bootstrapper.Test.LoaderFuncCAOTest.AppFabPushAndRetrieveData() in C:\SVN\2.3_Drop3\app\Src\Anz.Cre.Pdc.Bootstrapper.Test\LoaderFuncCAOTest.cs: line 281

Essentially what we are trying to do is the following.
we have different kinds of objects in our baseline. Objects of type CAO, Exposures, Limits that change everyday after close of business day.
We wanted to push these different objects in to respective named regions in the cache.
Region Name Objects
CAO ienumerable<caos>
Exposures ienumerable<exposures>
Limits ienumerable<limits>
we have a producer that pushes this data in to the cache and consumers of this data acting on the data when its available.
Now the issue we are facing is when we try to push around 4000 cao objects (roughly in the size of 300MB when serialized using xml) ,we are getting the above error. Increasing the size on the client and cache cluster didnt help.
The other alternative we were thinking about is chunking and pushing because appfabric doesnt support streaming. We might be able to push this data successfuly if we chunk. But how about the consumers ? wouldnt they face the same memory issue when we use
getallobjectsinregion ?
We thought if there was a way to figure out the keys in the region then probably the consumers can get one by one. However there is no such API.
The only option i see is using Appfabric notifications which msdn says isnt a reliable way.
Please help.

Unable to generate SQL trace file in shared mode ??

Hello,
I have a 10.2.0.1.0 database on Win2003 server 64 bits, with several instances.
On instances in dedicated mode, I can create trace files by using dbms_system.set_sql_trace_in_session, but on instances in shared mode nothing happens : no file is created in the udump directory.
This doesn't make any sense to me o_O
The shared parameters are the following :
dispatchers=(protocol=TCP)
shared_servers=1
I've checked all the necessary parameters to create a trace file (user_dump_dest, timed_statistics, max_dump_file_size...) and everything is fine.
I really don't understand why I can't create trace files when the instance is in shared mdoe...
Can anyone help about this weird issue ?
thank you !
regards,
Antoine

Remember also that in Shared Server mode the trace information is not written to a single file but is written to the Shared Server trace file. Before Oracle 10g there was no way to pull all the trace information for a single user session together out of the trace files. There is a tool for doing so with 10g.
HTH -- Mark D Powell --

PSE 10 will print from the organizer and in sharing mode, but not from full edit

I'm using pse10 on a computer with an i7 processor with 12GB of ram and windows 7 with a color laser printer. The print function works fine while in organizer and sharing mode, but not from full edit. When I try to print from full edit, I get an error message with a red "X" on it that says: "Before you can perform printer related tasks such as page setup or printing a document, you need to install a printer." I can't figure it out. The printer is in "ready" state, and works fine in the organizer mode, and in share. Any thoughts on what I'm missing? Thanks for any help in advance.

This occurs frequently with HP printers, though others are implicted as well from time to time
In Control panel>Devices and printers, right click your active printer, go to Printer Properties, and rename the printer to something short, e.g. my printer
http://kb2.adobe.com/cps/865/cpsid_86566.html

What files can be excluded from user-managed backup in noarchivelog mode?

what files can be excluded from user-managed backup in noarchivelog mode?
control files,
sqlnet.ora file,
tnsnames.ora,
datafiles associated with read-only tablespaces,
datafiles associated with read/write tablespaces,
of these, what can be excluded and what have to be included?
Thanks.

... and don't forget pfile or spfile... it can help too.
Re: What are generally backed up during the backup?
Nicolas.

Changing 'SYS' password in 'Shared' mode ?

Oracle 9.2/AIx5.3
I have two Db DB1/DB2, both of these Db's are in 'Shared' remote login password mode. Now, I was under the impression that whilst in this mode, it was not possible to simply issue a command
alter user SYS Identified by <NEW PASSWORD>;
to change the password, if you did, it would result in an error message. Today I have tried, and it allowed me to change the 'SYS' password ? How is this the case. I’ve tried it on a 10g Db that’s also running in 'Shared' mode and got the following error message.
ERROR at line 1:
ORA-28046: Password change for SYS disallowed

Metalink 452959.1 states you cannot change or add users to the passwordfile when remote_login_passwordfile='SHARED' and you need to set this to 'EXCLUSIVE'. I had to look this up and I found it simply by searching for the error number.
Assuming you are not running 10.2.0.1 you could have easily looked up yourself.
Also as Oracle is always changing, differences between 9.2 and 10gR1 or R2 are best verified by looking in the release notes for that version, providing the quickest answer easily.
Sybrand Bakker
Senior Oracle DBA

Starting WLS 10.3.5 Managed Server in MSI mode

(1) The first time you start the managed server in MSI, should the Admin Server be in Running Mode?
Or
(2) The Managed Server can be started in MSI mode (first time) with out having Admin Server in RUNNING MODE.
Which is true?
I think (1) is correct as the managed server directory wont be available under <DOMAIN_HOME>/Servers directory until you start the Managed Server. So only once you start the Managed Server, then only you can copy the config folder under Managed Server to start it in MSI mode (Having ADMIIN SERVER in SHUTDOWN state).
Is that correct?
So you can never ever start the Managed Server in MSI mode for the first time (with out having Admin Server in RUNNING MODE) after you create the managed server.
Is that correct?
Edited by: user734247 on Apr 18, 2012 2:58 PM

Your understanding is right.
The first time you are trying to start a managed server, you need admin server in RUNNING state.
From the next time onwards you can start managed server in MSI mode.
Although, as a workaround you can probably create certain files and directories to try and start managed server in MSI mode the first time itself:
1. Ensure that the entire domain directory is present in the machine where managed server is started.
2. Create directory <DOMAIN_HOME>/Servers/<ManagedServerName>/security
3. Create boot.properties file in <DOMAIN_HOME>/Servers/<ManagedServerName>/security/boot.properties with username and password
4. Create config-msi.xml copying from the config.xml in config directory
This list is not exclusive, it might need some more additional files.
You will have to create additional files like startServer.properties etc if we use node manager to start a managed server in MSI mode.
ORACLE strongly recommends to start the managed server through admin server the first time.
Arun

Draw on TiledImage as one unified image, still managing memory using tiles

what I am looking for is **exactly** this
http://www.garayed.com/java/186952-question-about-tiledimage-graphics2d-jai.html
"(...)How do I create an arbitrarily large TiledImage from scratch (lacking a
source image) and draw onto it using the TiledImage's graphics context
( TiledImage.createGraphics() ) without causing an
OutOfMemoryException?
The content of the image would be purely Graphics2D geometry. I don't
need to read in any external images, just to be able to create, and
draw on, very large blank images. I'm assuming that TiledImage is the
most appropriate class for this, but haven't been able to find any
sample implementations.
There must be some way to draw on a TiledImage as one unified image,
while still managing memory using tiles.(...)"
any ideas?
TIA

great article
my conclusion is that it seems java can't provide a huge graphics environment for drawing in a transparent manner, I think.
let's suppose I have a huge tiledImage. 20,000 x 20,000 pixels
now I want to draw a huge circle, radius = 10,000, center is the same square center
just an example, to clarify my doubt
when I request a Graphics2D context to draw this circle, tiledImage will intelligently and transparently request the necessary tiles to draw this circle, or I would have to code manually something to grab these tiles, calculate what to draw in each one and manage the tile cache?
TIA

Export memory using shared buffer

Hi
Lets say a user opens the PO screen ME22N in 2 separate windows accessing 2 separate PO numbers. If i use the export memory using shared buffer , how can i ensure that the data will not get mixed up ?
Any ideas?

YOu would have to get the session id to distinguish between the two. You can then use this id as part of your key to pass to the export statement.
Check this thread.
Quickest way to retrieve modeinfo[n],context_id_uuid from an ABAP pgm
Regards,
RIch Heilman

I accidentally set a remote mac to "remote management" while screen sharing. Can I remotely reset it to screen sharing?

I accidentally set a remote mac to "remote management" while screen sharing. I did not realize that this set it to require the commercial Apple Remote Desktop application to get to it. Can I remotely reset it to screen sharing? It has disconnected me from accessing it using iCloud.

Good question, and I am not sure. I believe the remote login WAS checked, so if that is needed, I think that is a available, unless checking Remote Management unchecked Login.
I am not computer illiterate, but I am not a professional IT guy. I read about more subjects than I know. Back to My Mac should still be set up in iCloud and I have seen some discussion of using ssh with that, and I am gathering that you are talking about ssh when you talk of remote shell access?
Thanks.

Advantage of createEnvironment in Shared mode

Hi,
Am looking for more information on what is shared when an environment is created in Shared mode.
We have a connection pool. For every task, a connection is made, statement is executed, statement is terminated and the connection is terminated. Next time the same task is executed, will there be any caching benefits from the previous execution?
Guessing that when you close the connection, all state is wiped out, hence there will be no advantage to the next guy?
Does creating an environment in shared mode keep some sort of statement cache?
Thanks much,
Nilofer

>>Am assuming I still need to do setStmtCacheSize(cacheSz).
All the Connections from a pool will inherit the statement cache size from the pool level setting. Everytime you get a session from a pool, it will have the stmt cache size set at the pool level. If you override that value by setting the stmt cache size on that Connection explicitly, tag that connection before you release it back to the pool, so that the stmt cache size value on that Connection is retained.
>>Do I also have to tag the statements? If I just do a terminateStatement(), since I have a statement cache, will the statement be automatically cached?
Yes. Tagging is not mandatory. It is useful for faster search.
>> On the same lines, if I do a createStatement(sql), will it return to me a statement matching the sql even though I don't provide the tag? (the documentation at http://download-west.oracle.com/docs/cd/B14117_01/appdev.101/b10778/reference008.htm#i1069559 suggests so)
Yes.
>>When should one tag the statements in a cache? To improve the search performance?
Yes.

Firewall Port for DHCP 2012 R2 Failover (Load Sharing mode)

Hi Everyone,
I was wondering if anyone can help me with finding a document for required Firewall Ports for DHCP 2012 R2 Failover (Load Sharing mode)
or just confirm if this is correct or not ?
TCP 647 for
DHCP failover messages between DHCP servers
TCP/UDP 67 and 68 initiate communication between the client and server
I am not sure if there is anything else
thanks in advance
Reza Negarestani

it was for a technical design document and I put this table for Firewall requirements what do you think ?
Direction
Port(s)
Bidirectional
TCP 647
Bidirectional
TCP 2535
UDP 2535
Bidirectional
TCP 67
TCP 68
UDP 67
UDP 68
Reza Negarestani

Shared-mode Event-driven capture @ 44.1 KHz -- GetBuffer() returning inconsistent and incorrect number of frames

I use WASAPI in shared-mode for render and capture in my application, and it works great with many different devices and sample rates from 8 KHz to 48 KHz.
I am experiencing issues with certain devices capturing @ 44.1 KHz, however. For instance, my Microsoft LifeChat headset works fine, but the onboard microphone for my dell laptop causes some trouble. When my capture thread wakes up from WaitForMultipleObjectsEx(),
I query for the NextPacketSize(). On some devices, I end up getting frames of 448 samples instead of 441, which completely throws off my logic as I am expecting integral size packets of audio. Then, after a few rounds of capturing 448 frames, I'll see a capture
of 378 frames (or something like that).
The device initializes properly (2 Byte PCM, 44.1 KHz) and returns S_OK. I am able to create an event and successfully set it with SetEventHandle(). My code is modeled after the examples for WASAPI capture. I'll post some sample code soon.

/* Main Audio Loop */
WaitForMultipleObjectsEx();
UINT nextPacketSize = 0;
hr = captureDevice.audioCaptureClient->GetNextPacketSize(&nextPacketSize);
if (FAILED(hr))
break;
while (nextPacketSize != 0)
UINT32 nRawFrames = 0;
hr = captureDevice.audioCaptureClient->GetBuffer((BYTE**)&pbData, &nRawFrames, &dwFlags, NULL, NULL);
if (FAILED(hr)) break;
// Handle flags
if (SUCCEEDED(hr))
// Resampling
... // need integral # frames here
// AEC and writing to buffer
// Cleanup
hr = captureDevice.audioCaptureClient->ReleaseBuffer(nRawFrames);
if (FAILED(hr)) break;
captureDevice.audioCaptureClient->GetNextPacketSize(&nextPacketSize);

Looking for values in memory in debugging mode

Hi,
I have the below submit statement in my program:
submit rm07mlbd
with werks eq w_werks
exporting list to memory and return.
How can I see the list exported to memory in debug mode?
As I donot have change authorisation, I cannot edit program to Display the list.

Double click on it in debugger?

Windows Server 2012R2 DHCP Load Sharing Mode

Hello,
I am planning to migrate a standalone DHCP Server on Windows 2003 to Windows Server 2012R2. With this migration, I like to use the 2012R2 Load Sharing Mode feature for DHCP. This implementation cover TWO sites.
My initial plan was to have two Windows 2012R2 DHCP Servers (virtual) at each site thus servers at each offers site redundancy as well. So two at Site 1 and two at Site 2 on load sharing mode. If ServerA at Site 1 goes down for some reason Server B
at Site1 will service the DHCP clients in load sharing mode without going to Site 2. I came across a MS document that says:
Number of Servers: DHCP failover is not supported for more than two DHCP Servers. The failover relationship is always comprised of two DHCP servers.
Does this mean I can have only TWO DHCP Servers in a load sharing mode. One at Site 1 and the second one at Site 2??
Thanks in advance
TIA TP

Hi,
The DHCP failover relationship is comprised of DHCP server and its partner. In a relationship there are just 2 DHCP servers.
But this doesn’t mean you can just use 2 DHCP servers in your sites. The DHCP failover setting is based on DHCP scope, for one DHCP scope there could be just one DHCP failover relationship including 2 DHCP servers. If you have more than 1 DHCP scope,
for example, set DHCP failover relationship between DHCP server 1 and DHCP server 2 for scope A. then set DHCP failover relationship between DHCP server 1 and DHCP server 3 for scope B.
The document Step-by-Step: Configure DHCP for Failover(http://technet.microsoft.com/en-us/library/hh831385.aspx) is a introduction about this function and for your reference.
Depends on the DHCP scope’s configuration and the requirements of your network. You may enhance or customize the architecture of implementation.
Best Regards,
Eve Wang

Managing memory in shared mode

Similar Messages

Maybe you are looking for