As DB2 Connect 10.5
FP5 goes GA, there are several exciting new capabilities to look
forward to. I want to highlight our emphasis on enhancing serviceability of DB2 Connect drivers for faster problem resolution.
Adoption of DB2 Connect has proliferated as customers strive to
modernize their applications for a distributed environment. Most of
these customers run with a complex multi-tiered stack environment
with several Application Servers and Databases. When an application
failure happens, it is often extremely difficult to pinpoint the
cause of the failure due to lack of sufficient diagnostics.
Although DB2 Connect drivers offer rich
tracing capability, collection of traces require recycle of
Application Servers, which is difficult in production environments.
Problem is aggravated if symptom is intermittent and workload is
high, leading to longer turnaround time for resolution and
frustration for all stakeholders (customer, IBM). To overcome this,
DB2 Connect Java driver now offers an online tracing capability,
which means that traces can be turned on/off and levels of tracing
can be altered without needing to bring down the Application
Server/JVM. An efficient polling mechanism is introduced whereby the
tracing related properties such as TraceFile and TraceLevel are
monitored at periodic intervals – these properties can be
dynamically altered and will be picked up during polling. This will
alleviate customer's pain point of bringing down application servers
just to collect traces in 24X7 production environments.
Another area we addressed was lack of
First Failure Data Capture. Again, we rely heavily on traces and
there was no way to collect minimal targeted diagnostic information
on a failure without needing to turn on traces (which can be heavy
weight). With Cancun, we have introduced the capability to collect
relevant diagnostics pertaining to specific error scenarios in both
Java (JDBC) and non-Java (CLI/ODBC) drivers without turning on
traces. When certain exceptions happen (which is configurable by
customers), basic diagnostic information on that exception is auto
logged in designated directories (which is also configurable). This
prevents loss of critical diagnostic data on the first occurrence of
the failure and improves the chances of resolution on the first
failure itself rather than waiting for a recurrence.
We have embarked on this journey to
enhance overall serviceability of DB2 Connect to drive up customer
satisfaction and improve overall product experience of all
stakeholders (application developers, App Server admins, DBAs,
systems programmers, IBM support and dev). There are several
serviceability related requirements on our radar for future releases.
If you would like to see a particular serviceability feature built
into the product in the long term, do drop me a note at
pallavipr@in.ibm.com.
No comments:
Post a Comment