By using this site, you agree to our updated Privacy Policy and our Terms of Use. Manage your Cookies Settings.
437,851 Members | 1,729 Online
Bytes IT Community
+ Ask a Question
Need help? Post your question and get tips & solutions from a community of 437,851 IT Pros & Developers. It's quick & easy.

HADR Questions

P: n/a
Anyone running HADR using VERITAS Cluster Server (VCS)? Our production
outsourcing company has recommended to use VCS instead of Tivoli System
Automation (TSA) for heartbeat monitoring on Linux.

Another question. We assume that automatic client reroute can be used with
Type 4 drivers for J2EE, and that we the type 4 driver version that is
distributed with DB2 Release 8.2. Is that correct?

What is the typical delay that people are experiencing for automatic client
reroute in a moderately high volume OLTP system?
Nov 12 '05 #1
Share this Question
Share on Google+
1 Reply


P: n/a
Another question. We assume that automatic client reroute can be used with
Type 4 drivers for J2EE, and that we the type 4 driver version that is
distributed with DB2 Release 8.2. Is that correct?
I believe that DB2 provides a "Universal" JDBC driver that provides
type 4 function and which is the recommended driver.

FYI - It has some additional client reroute capability beyond that
provided by just the base engine ACR feature. In particular, the
client can set up alternate server info prior to the first connection
attempt. This is valuable if you have new clients coming at the system
after a failover, because the base engine ACR feature can only set up a
client with alternate server information when that client actually
makes a successful connection.

For more info, search for "Rerouting DB2 Universal JDBC driver clients"
here:
http://publib.boulder.ibm.com/infoce...e/r0012130.htm
What is the typical delay that people are experiencing for automatic client
reroute in a moderately high volume OLTP system?


I don't have a direct answer to this question, but can add some color:

- This can depend on the TCP keepalive setting for the system, which
can shield the ACR facility from seeing certain kinds of connection
failures until the configured timeout has passed. If you want a fast
response, you'll want to tune down this setting.

- Aside from external factors like that, ACR is very responsive. When
deployed with HADR, it typically will be ready to reconnect much before
an HADR standby can finish taking over as primary.

Regards,
-Steve P.
---------------------------------
Steve Pearson
IBM DB2 UDB for LUW Development
Portland, OR, USA

Nov 12 '05 #2

This discussion thread is closed

Replies have been disabled for this discussion.