Are there any known issues with the DB_REGISTER support in 4.4.16?

I'm using Berkeley DB in a multi-process configuration (one
long-running writer, multiple short-running readers). When I activate
DB_REGISTER, it returns DB_RUNRECOVERY far too often, even though I'm
pretty sure that the reader processes properly close their environment
handles, and all processes accessing the environment set DB_REGISTER,
and no process opens the environment more than once.

Apart from that, DB_REGISTER adds a measurable delay when opening the
environment. To some extent, this was expected, but I'm surprised
that the delay is quite noticable.

There's a known problem and we'll be releasing a patch release of 4.4
soon to address it.

In short, there's a bug where DB_RUNRECOVERY is being returned too
often.

Here's a patch to apply in the meantime.

*** env/env_register.c.orig 9 Dec 2005 18:05:26 -0000 1.19
--- env/env_register.c 11 Dec 2005 16:12:18 -0000 1.20
***************
*** 216,224 ****
if (nr != PID_LEN)
goto corrupt;

! if (FLD_ISSET(
! dbenv->verbose, DB_VERB_REGISTER) &&
PID_ISEMPTY(buf)) {
! __db_msg(dbenv, "%02u: EMPTY", lcnt);
continue;
}

--- 216,224 ----
if (nr != PID_LEN)
goto corrupt;

! if (PID_ISEMPTY(buf)) {
! if (FLD_ISSET(dbenv->verbose,
DB_VERB_REGISTER))
! __db_msg(dbenv, "%02u: EMPTY", lcnt);
continue;
}

Regards,
--keith

=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=
Keith Bostic bostic (AT) sleepycat (DOT) com
Sleepycat Software Inc. keithbosticim (Yahoo IM)
118 Tower Rd. +1-781-259-3139
Lincoln, MA 01773 http://www.sleepycat.com

Thanks, it seems that this patch fixes the problem I was experiencing.