Hi,

Anyone know the fountions really mean?

"thread_id" founction --> DB_ENV->set_thread_id
"isalive" founction --> DB_ENV->set_isalive

Are they Create the thread we will use? Or they are only used to check
the status of the thread?

How to construct the "thread_id" and "isalive" fouctions?Could anyone
give me an simple example?

Thanks~

The thread_id function defaults to the right thing on systems that
support pthread mutexes. So in most cases you need not write your own
version. The thread_id function is only used for notating what thread
created a transaction or holds a mutex and is not required for proper
functioning of the BDB library unless you use the DB_ENV->failchk
interface. The is_alive function is used by the failchk interface to
determine if a thread is still alive. Please see:
http://www.sleepycat.com/docs/api_cxx/env_failchk.html

Michael Ubell
Sleepycat Software.

Thanks for your reply! Michae,

ubell (AT) sleepycat (DOT) com wrote:
I've looked through the various documentation on set_isalive, and
env_failchk. What I can't seem to figure out is how to write my
is_alive function. I'm using the CDS product with a number of
processes. Are each of them supposed to implement the is_alive function
or is only the process that calls failchk supposed to? Since my
processes are unrelated, is my is_alive suppose to somehow manage my
various processes pids? My processes are kicked off via a web server as
cgi programs. They don't live long.

I looked at DB_REGISTER but that doesn't seem to work for CDS.