wbratton 120000E4TY Visits (899)
Expert Advice: InfoSphere Streams ODBCDBG Failed to set isolation level
If you receive "ODBCDBG Failed to set isolation level" when trying to connect InfoSphere Streams to an Oracle database. This problem occurs because the Streams default isolation level for the query element is "REA
ubektas 27000478GM Visits (898)
kdillon 20000024FQ Visits (898)
KateDawson 060001TQ2C Visits (895)
As 2012 winds to a close (wow... where the heck did the year go?!?!?) we would like to thank all of our readers for a fantastic 2012! Our team of Support experts really enjoyed bringing you over 130 blog posts about top technical support content for all of our Information Management Products. We hope that you have found our blogs interesting and useful. We've got a number of ideas already for 2013... we'll be back bigger and better than ever!
Here are a few easy ways to find our blogs:
-- Your Thoughts from IM Support Team
100 Technical Support Tips -- posted between June 2011 and May 2012
wbratton 120000E4TY Visits (890)
James Kobielus, Senior Program Director for IBM's Big Data Analytics, is interviewed by Scott Laningham. During the interview Mr. Kobielus discusses why big data is so important, the role of Apache Hadoop and IBM BigInsights in making sense of big data, the evolving role of the data scientist, and and where data warehousing and big intelligence fit in.
ubektas 27000478GM Visits (886)
0XST_John_Gera 2700020XST Visits (883)
Using Virtual IP (VIP) provides a more elegant solution for Automatic Client Reroute between HADR Primary and Standby servers. The client does not have to be aware of the IP address of the Primary or the Standby server and can connect to the database only via the VIP. In cases where there have to be changes made to the Primary and Standby IP address/hostname, the client is still able to connect to the Primary using the VIP. Here is a demo showing VIP in action
Resource Info: Rollup Patch 1 for for 8.7 FixPack1 Metadata Asset Manager and Metadata Integration Bridges have been released
ubektas 27000478GM Visits (874)
Rollup Patch 1 for Metadata Asset Manager, Information Server Version 8.7 Fix Pack 1 is available at IBM Support Fixcentral site
Rollup Patch 1 for Metadata Integration Bridges, Information Server Version 8.7 Fix Pack 1 is available at IBM Support Fixcentral Site
ubektas 27000478GM Visits (873)
IBM InfoSphere Information Server, Version 9.1 fix list can be accessed at the link
That shows the summary of fixes included in the IBM InfoSphere Information Server, Version 9.1 release
There are also InfoSphere Information Server Version 9.1 fixes that require post-installation steps at the link
That shows certain APAR fixes require post-installation steps to use the fixes
amccarl 100000AY2G Visits (867)
TAG, YOU'RE IT!
We listen to your feedback - we really do!
In response, we've begun implementing Rating and now Tagging with our content. Here are some updates as to what's going on now:
* The TAGGING pilot has just begun (example: http
SEE??? We heard that change was desired, and here it is.
We're here listening and working to give you a better experience!
ubektas 27000478GM Visits (863)
IBM InfoSphere Information Server MustGather documents aid in problem determination and save time resolving Problem Management Records (PMRs).
You can gather the necessary information to improve the accuracy and completeness of the data collection process. This gathering of information can be done as follows:
* Automated gathering - Use the ISALite for InfoSphere Information Server collectors to gather log files and other artifacts. Refer to the link for automated information gathering.
* Manual gathering - Contact your IBM customer support representative before you gather information as outlined in the MustGather link for IBM InfoSphere Information Server document.
ubektas 27000478GM Visits (861)
0XST_John_Gera 2700020XST Visits (841)
There are situations where in a HADR environment, transaction on the Primary are bloc
Brewer's CAP Theorem:
Brewer's CAP theorem states that, in a Distributed Computer System, we can guarantee only two of the following simultaneously:
- Consistency (all nodes see the same data at the same time)
- Availability (a guarantee that every request receives a response about whether it was successful or failed)
- Partition tolerance (the system continues to operate despite arbitrary message loss or failure of part of the system,)
Figure 1. Venn Diagram above shows that there can be only 2 of Consistency(C), Availability (A) or Partition tolerance(P) at a given time
Now putting that in terms of DB2 HADR for a LOGGED OPERATION:
Consistency (C) - Data is the SAME on Primary and Standby
Availability (A) - Clients are ALWAYS able to connect and get a response (i.e no blocking)
Partition Tolerance (P) - Netw
Basically the theorem says that in the presence of a partition event (like a network outage between the 2 hadr nodes) there MUST BE a tradeoff between availability and consistency.
HADR sync modes are the primary dial to control C vs. A tradeoffs:
Transaction are committed on the Primary after relevant logs have been written to disk on Primary and:
SYNC --> logs have been written to disk on Standby as well ==> Highest C (Data guaranteed on Standby)
NEARSYNC --> Logs are received into memory on Standby ==> More C
ASYNC --> Logs sent to the Standby on the Network (Don't wait for ACK) ==> More A
SUPERASYNC --> Don't wait for Standby!! ==> Highest A (No guarantee)
In SYNC mode, the total time for a log write = Σ (Primary_Log_Write + Log_Send + Standby_Log_Write + Ack_Message)
Figure 2: Shows the various sync modes
Further fine tuning C vs A :
HADR_TIMEOUT --> How long to wait before considering communication failure. Till we reach this time we are still considered PEER ==> More C
HADR_PEER_WINDOW --> In case of communication failure, how long to wait before actually breaking the connection ==> More C
When both HADR_PEER_WINDOW and HADR
Where does DB2_HADR_BUF_SIZE and DB2
DB2_HADR_BUF_SIZE --> For increasing the HADR standby log receive buffer size
Higher DB2_HADR_BUF_SIZE and DB2
In situation where we see the Transaction being blocked on the Primary, it is because the HADR setup is in favor of C vs A. For example:
1. Sync Mode = SYNC/NEARSYNC
2. There is a Partition event (standby log buffer full/network failure)
Then we see blocking on Primary i.e no A. In such situations we can tune:
- DB2_HADR_BUF_SIZE - (to get CA)
Note: Starting v10.1 we have introduced a new parameter 'HADR_SPOOL_LIMIT'. Log data that is sent by the primary is written, or spooled, to disk on the standby if it falls behind in log replay. The standby can later on read the log data from disk.
Bada Bing Bada Boom!! :)
wbratton 120000E4TY Visits (837)
Expert Advice: Packaging Jaql modules inside an applications for a Cloudera overlay environment
When the InfoSphere BigInsights 1.4 server has been installed over a Cloudera cluster, an application that uses Jaql modules will need to package the module with the application. Click here for more information...
amccarl 100000AY2G Visits (832)
A new development community has been launched for IBM software customers submitting enhancement requests for products. There you have an opportunity to collaborate directly with the IBM product development teams and other product users. The discussions are free and public. You can view the Top 20 watched or voted Feature Requests, along with those that are submitted, planned, or delivered.
See it here: http
Note: Currently, only DB/2 for z/OS, DB2 Utilities Suite, DB2 Sort for z/OS, and DB2 Automation Tool for z/OS are supported.