Informatica 7.1.4: IDLE Connections Killed

So I posted this question on Informatica's forum "Communities" but I wanted to post it here as well.
We're on Informatica 7.1.4 (yeah, I know its old, but we're about to upgrade) and we have been running into issues where one long running ETL fails because the Oracle 11gR1 DB terminated it's IDLE connection. From what I can understand, here's what happens.
1. We have a max concurrency rate of 10. So suppose 9 other ETLs are running along with our 1 long running ETL (1hr25minutes is the average time for this one).
2. All the 10 ETLs create DB connections for the Writer thread at the kick off of the ETL
3. The long running ETL takes an hour before it even gets to the point where it can start writing to the target. So basically it's connection has been IDLE the entire 1 hour.
4. On the DB side of things, the DBAs have a cleanup script to remove IDLE sessions that haven't done anything in 1+ hours. So their script kills the connection that the Writer thread is about to use.
5. Our long running ETL finally finishes sourcing and processing the data and it is ready to insert, but now it fails because the connection has been terminated by the DB.
I know the easy out is to have the DBAs not kill connections made by Informatica, but I was hoping there would be a more elegant solution.
Is there a way to have Informatica revalidate it's connection prior to writing? Something like, if connection doesn't exist, re-create, else write? It seems like this should be a server level setting.
Or is there a way to have Informatica send some kind of heartbeat to the Oracle DB to keep the connection alive?
Right now we're definitely having the DBAs avoid killing connections created by Informatica. But for the long term, we would like to see Informatica be more intelligent about the health of the connection its using.
Thanks for the advice.
-Joe

Are both of the Informatica Servers on the same machine?
If so, one thing you can try is when connecting to the Remote Server from Windows is setting the IP Addresses to 127.0.0.1 or using the name localhost for the Informatica Server (In the Server Editor screen of the Workflow Manager). You only need to set the valid IP Address for the Repository Server to get the Repository Admin tool to connect. Once that connection is made the Informatica Server uses that information to start itself.
I used an approach like this to stand up both Informatica Servers on Oracle Enterprise Linux.

Similar Messages

Stale connection kills the VM

Here is the scoop: a process (JDK 1.3 w/ Oracle OCI driver v8.1.6, running on Windows 2000) establishes a connection to the database (Oracle 7.3.4) and opens a number of PreparedStatement(s). While the connection is open and idle, the database instance bounces. After that, any calls on PreparedStatements on that connection kills the VM with this message:
# An EXCEPTION_ACCESS_VIOLATION exception has been detected in native code outside the VM.
# Program counter=0x604b0ae1
# Problematic Thread: prio=5 tid=0x12738ee0 nid=0x514 runnable
I believe an error occurs in the native OCI driver code and is not mapped to a JVM exception.
Did anyone else see this? On a related subject, is there a good way to ping the conneciton? (isClosed() did not work)
null

Anyone? Bueler?

Huge number of idle connections from loopback ip on oracle RAC node

Hi,
We have a 2node 11gR2(11.2.0.3) oracle RAC node. We are seeing huge number of idle connection(more than 5000 in each node) on both the nodes and increasing day by day. All the idle connections are from VIP and loopback address(127.0.0.1.47971 )
netstat -an |grep -i idle|more
127.0.0.1.47971 Idle
any insight will be helpful.
The server is suffering memory issues occasionally (once in a month).
ORA-27300: OS system dependent operation:fork failed with status: 11
ORA-27301: OS failure message: Resource temporarily unavailable
Thanks

user12959884 wrote:
Hi,
We have a 2node 11gR2(11.2.0.3) oracle RAC node. We are seeing huge number of idle connection(more than 5000 in each node) on both the nodes and increasing day by day. All the idle connections are from VIP and loopback address(127.0.0.1.47971 )
netstat -an |grep -i idle|more
127.0.0.1.47971 Idle
any insight will be helpful.
The server is suffering memory issues occasionally (once in a month).
ORA-27300: OS system dependent operation:fork failed with status: 11
ORA-27301: OS failure message: Resource temporarily unavailable
Thankswe can not control what occurs on your DB Server.
How do I ask a question on the forums?
SQL and PL/SQL FAQ
post results from following SQL
SELECT * FROM V$VERSION;

ASA TCP Idle Connection Timeout Suspense

Hello I upgraded our Cisco ASA 5520 with a Cisco ASA 5585. Though both ASA were configured with default TCP Idle Connection Timeout values people are now starting to complaint that idle SSH connections are being terminated. This is proper behavior but they were claiming it didn't occur with the old firewall. Our users are setting keepalives for 1800 seconds to get around this before I can bump the setting to infinite (setting 0). Is there a bug with the feature in older ASA OS?

Hi,
Before looking for a bug I would check the ASA logs (hopefully you are storing them to a separate Syslog server) and see why the connections are torn down (Teardown reason) and how long have they been on the ASAs connection table before they were torn down.
You also have the option to perform traffic capture on the ASA for the traffic in question and confirm why or which party terminates the connection.
I guess you can use the MPF on the ASA to configure separate idle timeouts for just these SSH Connections if you do not want to touch the global timeout values.
I have not run into any problems with the timeout settings on the older softwares. In the newer softwares (8.3+) I have run into these problems. In those situation the ASA has not removed the connection that have reached the timeout value. I have seen connection that have been idle for over 1000h.
- Jouni

ASA-5585-X 8.4(6)5 Idle connections are not being removed according to timeout settings

Hi,
Just a quick question if anybody has run into a bug where the ASAs "timeout" settings are not being applied to idle connections.
It seems that our ASA running the software level 8.4(6)5 is not tearing down connections. This mainly seems to be a problem in one Security Context where there are around 300k UDP connections (related to VOIP phones) that are not being torn down. Idle timers on the connections are going as far as 700 hours. Common to all the UDP connections is also the fact that only 19 Bytes of data has been transmitted on the connection built on the firewall. I am not sure what the purpose of these UDP Connections is as both the source and destination port is a random high port.
I was not able find any Bug ID which description would match the situation I am seeing. I did not see anything in the release notes of 8.4(7) or its interrim release either that would list thing kind of bug.
- Jouni

Hi Jouni,
This caveat seems to be the closest match as 8.4.6 is the affected ASA code.
CSCuh13899
Symptoms:-
Some connection may not removed even after reaching idle timeout.
https://tools.cisco.com/bugsearch/bug/CSCuh13899/?reffering_site=dumpcr
You can upgrade to the next stable ASA code as suggested in the referred document.
HTH
Regards,
Dinesh Moudgil
P.S. Please rate helpful posts.

EclipseLink timeout on idle connections

Hi,
I'm using EclipseLink 2.1.0 + MySql for a web application. Internal connection pool is used instead of Tomcat maintained pool.
The issue I'm currently facing is that connections in the pool become idle during the night (when the application is indeed idle).
In the morning the first attempt to use a connection fails, but further attempts work fine.
I've heard that H3C0 (Hibernate) and other connection pools support a feature that sends a sort of heartbeat packet to
keep the connection alive so that every connection in the pool does not timeout (e.g. using the testConnectionOnCheckout
property).
Is it possible to get the same behaviour on EclipseLink in order to avoid that an idle connection times out ?
Matteo

It turned out that the feature I was looking for is not implemented in EclipseLink 2.1.0.
In order to fix my problem I had to change the configuration of EclipseLink to use the Tomcat connection pool (dbcp). This provides the following useful parameters:
     testWhileIdle="true"
     validationQuery="select count(*) from dual"
     minEvictableIdleTimeMillis="2880000"
Basically this checks every 8h idle connections performing the validation query, thus avoiding timeout. Simple and good!

SRP541W drops idle connections fast

I have installed SRP541W and found a big anoying issue with this router - it drops idle TCP connections after a couple of minutes (more than 5 but less than 20) of inactivity. Namely, this "feature" becomes apparent when using ssh connections - if nothing happens in an ssh terminal for a given time, the connection is lost and the ssh client must be reconnected. As having a lot of ssh terminals open simultaneously (and, of course, leaving some them idle for a while) is a core of my work, this behavior annoys me a lot. Does anyone know how to solve this problem?
I understand that dropping idle connection is needed to keep the NAT tables reasonably small, but dropping connections after less than 20 minutes of inactivity is just way too fast and useless. Even cheap home routers usually do not drop idle connections after many hours or days of inactivity.

Also, if I switch the Airport Express to connect to the 2.4GHz network instead, the same thing happens on that network. After a few hours, the base station stops authenticating clients on the 2.4GHz network, and only the 5GHz network works.

DPS 6 idle connection timeout?

Ok I must be blind... where is the idle connection timeout setting for DPS 6? I would assume in the connection handler. Migrating from v5 and in the old version its in the network group.

Is this what you're looking for?
$ dpconf help-properties | grep -i idle
ldap-data-source monitoring-inactivity-timeout rwd 1s <= DURATION | unlimited Maximum time the availability monitor detects no activity before performing a search on the idle connection to keep it alive. (Default: 2m)
ldap-listener connection-idle-timeout rwd DURATION | unlimited Maximum time a client connection can remain idle before being closed (Default: 1h)
ldaps-listener connection-idle-timeout rwd DURATION | unlimited Maximum time a client connection can remain idle before being closed (Default: 1h)

Hi All, I am trying to install informatica 9.1 HF 2 connecting to 11g DB. when installing Informatica, I am at the DOMAIN CREATION windows and inputting the DB connect details and for some reason, it cannot connect to the DB. The DB is running any id

Hi All,
I am trying to install informatica 9.1 HF 2 connecting to 11g DB.
when installing Informatica, I am at the DOMAIN CREATION windows and inputting the DB connect details and for some reason, it cannot connect to the DB.
The DB is running
any ideas?
the error message states:
THE CONNECTION FAILED: CORRECT THE DATABASE INFO AND TEST THE CONNECTION AGAIN..any help please
thanks

Yup, that did it.
I actually had been made aware of the patch when I downloaded CF, but for some reason I thought it was difficult to install (I thought that you could only use the installer package if you could open Administrator). IAC I downloaded the appropriate patch file, ran it and Administrator ran. There was something in the install notes about reconfiguring the websites and removing the IIS6 Management feature but I'm not going to fool with anything because right now at least CF is running.
Thanks.

Difference between Idle Connection Time and HTTP Keepalive Duration?

What is the difference between the Idle Connection Timeout and the HTTP Keepalive Duration configuration parameters? I thought that setting one or both of these values would cause the webLogic web server to terminate client requests that result in processes that run much longer then the time value set in these parameters.

Both of these settings are protection against denial of service at the protocol level, Brian.
Idle Connection Timeout is a generic timeout for idle connections, whereas HTTP Keepalive Duration is specifically for HTTP keepalive connections. (which doesn't cover T3, IIOP or jCOM)
Both of those configuration options are related to freeing up resources like sockets and server-side objects in memory from idle clients.
There's actually no way of stopping long-running processing on the server triggered by a client request that I'm aware of - there are ways you can alert system administrators that something is wrong however.
The best bet there is defining a value for Stuck Thread Max Time, and also for Stuck Thread Timer Interval. That won't stop processing on the server, but it will cause a log message that can be sent on to a monitoring console like Openview.
Beyond that, it's really an issue for application development. There are ways that you can put the brakes on, like ensuring that poentially long-running operations run under a transactional context, and setting a reasonable transaction timeout value. Also, checking for rollback-only settings inside EJB code.
I could be wrong, but I'm pretty sure there's no easy answer here.
Does that help at all, Brian?
 
Kevin Powe
http://www.integral-techsolutions.com

Error occuring during manual login after idle connection time out

Hi,
I am using a PowerBuilder (v11.1) application to access Crystal Reports (BOXI Enterprise v 3.1) using Opendocument. The app server is Apache Tomcat. We have implemented SSO using Vintela and at present we have set idle connection time out to 1 hr. The problem I am facing is listed below:
The SSO Authentication works fine for accessing these reports. But when we try to run crystal reports after not using it for more than 2 hours system prompts us for login. I understand that we are getting login prompt because connection was idle for more than 1 hour but when we manually enter our login credentials it gives us an error 'An error has occured. An error occured while trying to view the document'.
However after getting the above error if we again try to run the report, it works fine.
Please can anyone provide us with some insight as to why this behaviour could be occurring. Any help would be much appreciated.

->in application connection time out is set to zero
Are you sure? There are two different can timeouts you can experience: connection timeouts and command timeouts. The link posted by Praveen is about command timeouts. The default for the connection timeout is 15 seconds and the command timeout is 30 seconds.
It is very reasonable to set the command timeout to 0, which means that wait forever. Waiting forever is less reasonable for the connection timeout. But when I think of it, I am not sure that 0 means "forever" for the connection timeout. So if you
have set it to zero, maybe that is your problem.
Else, I would suspect a network problem.
Erland Sommarskog, SQL Server MVP, [email protected]

Avoiding the idle connection timeout?

Our DBAs have implemented an idle connection timeout.
Is there an option in SQL-Developer to keep the connection alive by having it periodically send "dummy" commands to keep the connection alive?

Hi,
Even though there is no listing for it on the Third Party Extensions Exchange on the SQL Developer Community page here,
http://www.oracle.com/technetwork/developer-tools/sql-developer/community/index.html
a "Keep-Alive" extension exists and you may wish to try it. Here is the post for it on this forum:
SQL Developer Keep-Alive Extension Created!
If you don't wish to install an extension, an alternative is here:
How to reconnect without using the "keep alive" plug-in?
Regards,
Gary
SQL Developer Team

Multiple IDLE connections

Hi,
I have two IDLE enabled IMAP accounts and I'm trying to work out if everything is working properly as it doesn't feel like it is.
When I check netstat it shows multiple established connections for each account which seems to many for it to be right.
Any thoughts on what to check? Server is showing two connections which would be right.
Thanks,
Adam

If you must have a 1-1 app user/connection to database user/connection, then using a DatabaseSession would probably be the easiest solution. The means that each user would create and connect to their own DatabaseSession when they login to the app. You could still share the same TopLink project instance to reduce some of the connecting/initialization overhead, but there would be no way to pool connections. Using a DatabaseSession would not allow for any shared caching across users.
Do you need to be connected as the user for reading, or just for writing to the database? If you can use a shared user/connection for reading, then you could use a ServerSession and have a shared cache. Define the shared user in the ServerSession, and when each user connects acquire a ClientSession through the acquireClientSession(DatabaseLogin) API, which allows you to specify the user/password to connect as for writing.
In TopLink 10.1.3 you may also wish to investigate the VPD and exclusive client connection support. This allows for some of your data to be read through a user/VPD secure connection, and other shared data to still be cached. It also allows for the user/VPD to be switched on a pooled connection.

Informatica Workflow not able to connect to source database.

Hi,
I have completed the installation of OBI Apps.All the test connections are working fine and I have also configured the Relational Connection in Informatica Workflow.The passwords,username and connect string are correct.Still when I run a ETL all the tasks that need to connect to source database fail.I checked the session logs of those workflow and it gave me following error:
READER_1_1_1> DBG_21438 Reader: Source is [UPG11i], user [obiee]
READER_1_1_1> CMN_1761 Timestamp Event: [Fri Sep 05 18:01:37 2008]
READER_1_1_1> RR_4036 Error connecting to database [
Database driver error...
Function Name : Logon
ORA-12154: TNS:could not resolve the connect identifier specified
Database driver error...
Function Name : Connect
Database Error: Failed to connect to database using user [obiee] and connection string [UPG11i].]
From DAC I am able to connect to the databases.There seems to be some problem in the relational connections.What are the drivers involved and where should they be installed.
Please help.
Thanks and regards,
Soumya.

Hi,
I think you need to check you connections in the Work flow manager. So, i'd suggest you to follow this below section again. Your error clearly shows that there is problem with the connections.
"4.13 Configuring Relational Connections In Informatica Workflow Manager" in Installation & Configuration guide.
Post back here, if you find any problems.
Thanks,

Informatica Workflow Manager ODBC Relational Connection for ETL in DAC

In Informatica Workflow Manager, I have created a Relational Connection of type ODBC and specified Connect String as "DSN=BIEEDW" where "BIEEDW" is the System ODBC DSN already set pointing to a SQL Server 2008 database.
However, when the ETL run in DAC, the following error occurs in Session log files showing that the database and driver cannot be located:
MAPPING> CMN_1569 Server Mode: [UNICODE]
MAPPING> CMN_1570 Server Code page: [MS Windows Traditional Chinese, superset of Big 5]
MAPPING> TM_6151 The session sort order is [Binary].
MAPPING> TM_6156 Using low precision processing.
MAPPING> TM_6180 Deadlock retry logic will not be implemented.
MAPPING> TM_6187 Session target-based commit interval is [10000].
MAPPING> TM_6307 DTM error log disabled.
MAPPING> TE_7022 TShmWriter: Initialized
MAPPING> DBG_21075 Connecting to database [DSN=BIEEDW], user [bieedw02]
MAPPING> CMN_1761 Timestamp Event: [Wed May 22 01:29:17 2013]
MAPPING> CMN_1022 Database driver error...
CMN_1022 [
[Microsoft][ODBC Driver Manager] Data source name not found and no default driver specified
Database driver error...
Function Name : Connect
Database driver error...
Function Name : Connect
Database Error: Failed to connect to database using user [bieedw02] and connection string [DSN=BIEEDW].]
Any hint in setting the Connect String for ODBC Relational Connection?

Hi,
Let me tell you the real story:
Our server architecture consists of two servers:
Windows Server 2008 R2 (64-bit) platform with the following installed:
- SQL Server 2008
- DAC 10.1.3.4.1
- OBIEE 11g
- BI Apps (Financial Analytics) 7.9.6.3
- Informatica Server 9.1.0 HotFix 2
Windows Server 2003 Enterprise Edition SP2 (32-bit) platform with the following installed:
- Informatica Clients (i.e. Workflow Manager, Repository Manager, Designer and Workflow Monitor)
And thus the ODBC Relational Connection is configured in Informatica Client machine (Workflow Manager) which is a 32-bit platform.
Any idea?

Informatica 7.1.4: IDLE Connections Killed

Similar Messages

Maybe you are looking for