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

RE: slapadd segmentation fault (2.1.22/bdb 4.1.25)



> -----Original Message-----
> From: owner-openldap-software@OpenLDAP.org
> [mailto:owner-openldap-software@OpenLDAP.org]On Behalf Of Allan Streib

> Re-ordering my ldif file (produced by slapcat) into superior order
> avoids this problem.  My speculation is that slapadd maintains some
> kind of forward reference when it encounters a "child" entry before
> encountering its "parent",

That's correct.

> and that with large ldif files (about 20
> million lines in this case, representing about 1.2 million entries,
> with up to 4 levels of "depth" from the root entry to the
> leaf entries) perhaps it is overflowing something.

The code is set to dynamically resize the list of forward references. Perhaps
you simply ran out of memory.

> I do have another question -- the data load produced over 8 GB of BDB
> log files, and db_archive did not say that any were removable until
> after slapadd finished.  Is this a result of my
> "DB_TXN_NOSYNC" option
> in DB_CONFIG, or is that normal?

That's normal. slapadd doesn't checkpoint the database while it runs, it only
does it before exiting. Since it's common to run it with DB_TXN_NOSYNC it
seemed counterproductive to checkpoint, but perhaps that should be changed.

  -- Howard Chu
  Chief Architect, Symas Corp.       Director, Highland Sun
  http://www.symas.com               http://highlandsun.com/hyc
  Symas: Premier OpenSource Development and Support