[Fedora-directory-users] Slapd Crash and wont start up again

[Noriko Hosoi]

Jim Summers wrote:
> Hello All,
>
> Had a power outage this morning.  When the system came back up the 
> slapd  would not start.  I found the following in the error log:
>
>         Fedora-Directory/1.0.2 B2006.060.1928
>         landin.cs.ou.edu:636 (/opt/fedora-ds/slapd-landin)
>
> [04/Jul/2008:16:17:37 -0500] - Entry "cn=SNMP,cn=config" -- attribute 
> "nsSNMPName" not allowed
> [04/Jul/2008:16:17:37 -0500] - Entry "cn=SNMP,cn=config" -- attribute 
> "nsSNMPName" not allowed
> [04/Jul/2008:16:17:37 -0500] - Fedora-Directory/1.0.2 B2006.060.1928 
> starting up
> [04/Jul/2008:16:17:37 -0500] - ERROR: Database version mismatch 
> (expecting 'Netscape-ldbm/7.0' but found 'nsibleObject' in directory 
> /opt/fedora-ds/slapd-landin/db)
> [04/Jul/2008:16:17:37 -0500] - start: db version is not supported
> [04/Jul/2008:16:17:37 -0500] - Failed to start database plugin ldbm 
> database
> [04/Jul/2008:16:17:37 -0500] - WARNING: ldbm instance userRoot already 
> exists
> [04/Jul/2008:16:17:37 -0500] - WARNING: ldbm instance NetscapeRoot 
> already exists
>
>
> Any ideas on how I can fix this?  This is a replica and fortunately 
> the master is ok.  Would it be possible to simply copy the db 
> directory from the master and drop it on this replica?
First, you may want to try fixing the corrupted DBVERSION files.

Please search DBVERSION files in your db directory.  There is one in the 
db dir, and one in each backend directory (e.g., userRoot).  Open the 
file and if the content is not "Netscape-ldbm/7.0", replace the content 
with the string.  If other db files are not corrupted, your replica 
would start without the database version mismatch error.  (You may still 
want to run replica initialization to make sure the replica is healthy.)

If it still does not start, you have to recover the replica.  As you 
asked, you could "copy" the master db to the replica, but there are some 
steps to follow.

If your master and the replica have the same backend configuration, you 
could copy the directory EXCEPT __db.* files.

If they don't, you could make a back up on the master with "db2bak", 
then restore on the replica server just the backend to be replicated.
http://www.redhat.com/docs/manuals/dir-server/cli/8.0/Configuration_Command_File_Reference-Command_Line_Scripts-Shell_Scripts.html#Configuration_Command_File_Reference-Shell_Scripts-bak2db_Restore_database_from_backup

Hopefully, it recovers your replica.

Thanks,
--noriko

-------------- next part --------------
A non-text attachment was scrubbed...
Name: smime.p7s
Type: application/x-pkcs7-signature
Size: 3237 bytes
Desc: S/MIME Cryptographic Signature
URL: <http://lists.fedoraproject.org/pipermail/389-users/attachments/20080705/b5dfff70/attachment.bin>