[Date Prev][Date Next] [Chronological] [Thread] [Top]

Re: alock + ldbm in 2.3.21

--On Wednesday, April 12, 2006 4:10 PM -0400 matthew sporleder <msporleder@gmail.com> wrote:

> Apr 11 15:21:22 labogldir02 slapd[16232]: [ID 658149 local6.debug]
> ldbm_back_db_open: alock package is unstable; database may be
> inconsistent!
> Apr 11 15:21:22 labogldir02 slapd[16232]: [ID 100111 local6.debug]
> slapd starting
> Can I get a status on alock for LDBM?
> According to:
> http://www.openldap.org/devel/cvsweb.cgi/servers/slapd/alock.c?hideatt
> ic= 1&sortbydate=0
> it looks like ldbm is having some alock rethinks.  Should I wait for
> ldbm to be re-stabilized in 2.3.x?

LDBM is going to be removed from OpenLDAP 2.4, and there are many reasons
not to use LDBM (stability, etc).  I would rethink why you are using LDBM
in the first place (not that the problem you are reporting doesn't need
to be addressed).

The reason I'm using LDBM is that I can migrate my servers to 2.3 -very quickly- by sticking with the same backend as I was using in 2.1. Basically, my procedure was: install binaries, run db_upgrade, start slapd. (takes about two minutes)

Is this error because my database directories did not contain the
'alock' file when I started up?  I saw that 'alock' was created when I

I can do some testing on this once my slamd run is finished tomorrow,
but I would appreciate some direction.

How large is your database?

If it is not absolutely huge, it should be fairly quick to load it using slapadd -q, and if you have a multiple CPU system, you can use the multi-threaded nature of slapadd in 2.3 to speed it up even more (see the tool-threads directive). This is all assuming you are using BDB of course.

I would really rethink using LDBM, especially since you are upgrading to a modern version of OpenLDAP.

Now that you've started slapd, and alock is created, does LDBM continue to complain, or are things happy?


Quanah Gibson-Mount
Principal Software Developer
ITS/Shared Application Services
Stanford University
GnuPG Public Key: http://www.stanford.edu/~quanah/pgp.html