I had the same exact problem on CentOS 5.2 yesterday. ns-slapd fails trying to open the status file read/write. The directory if owned by root, and it should be owned by ns-slapd's uid (nobody by default). I believe the directory is /var/run/dirsrv.


nh
On Sep 18, 2008, at 6:27 AM, David Carter wrote:

When I run either setup-ds.pl or setup-ds-admin.pl, I go through all 
the prompts and get an error saying it's unable to start the server 
and that I should check the error logs. Here's the contents of the 
error log:

[18/Sep/2008:11:46:35 -021800] - dblayer_instance_start: pagesize: 
4096, pages: 128743, procpages: 7693
[18/Sep/2008:11:46:36 -021800] - cache autosizing: import cache: 204800k
[18/Sep/2008:11:46:36 -021800] - li_import_cache_autosize: 50, 
import_pages: 51200, pagesize: 4096
[18/Sep/2008:11:46:36 -021800] - WARNING: Import is running with 
nsslapd-db-private-import-mem on; No other process is allowed to 
access the database
[18/Sep/2008:11:46:36 -021800] - dblayer_instance_start: pagesize: 
4096, pages: 128743, procpages: 7693
[18/Sep/2008:11:46:36 -021800] - cache autosizing: import cache: 204800k
[18/Sep/2008:11:46:36 -021800] - li_import_cache_autosize: 50, 
import_pages: 51200, pagesize: 4096
[18/Sep/2008:11:46:36 -021800] - import userRoot: Beginning import 
job...
[18/Sep/2008:11:46:36 -021800] - import userRoot: Index buffering 
enabled with bucket size 100
[18/Sep/2008:11:46:36 -021800] - import userRoot: Processing file "/
tmp/ldif20a8Hl.ldif"
[18/Sep/2008:11:46:36 -021800] - import userRoot: Finished scanning 
file "/tmp/ldif20a8Hl.ldif" (9 entries)
[18/Sep/2008:11:46:37 -021800] - import userRoot: Workers finished; 
cleaning up...
[18/Sep/2008:11:46:37 -021800] - import userRoot: Workers cleaned up.
[18/Sep/2008:11:46:37 -021800] - import userRoot: Cleaning up producer 
thread...
[18/Sep/2008:11:46:37 -021800] - import userRoot: Indexing complete.  
Post-processing...
[18/Sep/2008:11:46:37 -021800] - import userRoot: Flushing caches...
[18/Sep/2008:11:46:37 -021800] - import userRoot: Closing files...
[18/Sep/2008:11:46:37 -021800] - All database threads now stopped
[18/Sep/2008:11:46:37 -021800] - import userRoot: Import complete.  
Processed 9 entries in 1 seconds. (9.00 entries/sec)
[18/Sep/2008:11:46:38 -021800] - Fedora-Directory/1.1.2 B2008.248.1443 
starting up
[18/Sep/2008:11:48:06 -021800] - Fedora-Directory/1.1.2 B2008.248.1443 
starting up
[18/Sep/2008:11:52:46 -021800] - Fedora-Directory/1.1.2 B2008.248.1443 
starting up
[18/Sep/2008:11:53:31 -021800] - Fedora-Directory/1.1.2 B2008.248.1443 
starting up
[18/Sep/2008:12:06:02 -021800] - dblayer_instance_start: pagesize: 
4096, pages: 128743, procpages: 7693
[18/Sep/2008:12:06:02 -021800] - cache autosizing: import cache: 204800k
[18/Sep/2008:12:06:02 -021800] - li_import_cache_autosize: 50, 
import_pages: 51200, pagesize: 4096
[18/Sep/2008:12:06:02 -021800] - WARNING: Import is running with 
nsslapd-db-private-import-mem on; No other process is allowed to 
access the database
[18/Sep/2008:12:06:02 -021800] - dblayer_instance_start: pagesize: 
4096, pages: 128743, procpages: 7693
[18/Sep/2008:12:06:02 -021800] - cache autosizing: import cache: 204800k
[18/Sep/2008:12:06:02 -021800] - li_import_cache_autosize: 50, 
import_pages: 51200, pagesize: 4096
[18/Sep/2008:12:06:03 -021800] - import userRoot: Beginning import 
job...
[18/Sep/2008:12:06:03 -021800] - import userRoot: Index buffering 
enabled with bucket size 100
[18/Sep/2008:12:06:03 -021800] - import userRoot: Processing file "/
tmp/ldif8U2ox1.ldif"
[18/Sep/2008:12:06:03 -021800] - import userRoot: Finished scanning 
file "/tmp/ldif8U2ox1.ldif" (9 entries)
[18/Sep/2008:12:06:04 -021800] - import userRoot: Workers finished; 
cleaning up...
[18/Sep/2008:12:06:04 -021800] - import userRoot: Workers cleaned up.
[18/Sep/2008:12:06:04 -021800] - import userRoot: Cleaning up producer 
thread...
[18/Sep/2008:12:06:04 -021800] - import userRoot: Indexing complete.  
Post-processing...
[18/Sep/2008:12:06:04 -021800] - import userRoot: Flushing caches...
[18/Sep/2008:12:06:04 -021800] - import userRoot: Closing files...
[18/Sep/2008:12:06:04 -021800] - All database threads now stopped
[18/Sep/2008:12:06:04 -021800] - import userRoot: Import complete.  
Processed 9 entries in 1 seconds. (9.00 entries/sec)
[18/Sep/2008:12:06:04 -021800] - Fedora-Directory/1.1.2 B2008.248.1443 
starting up

It seems to think we've started. Not a lot of help. So I run 'sevice 
dirsrv start', which fails. The only new line in the error log is:

[18/Sep/2008:13:20:57 -021800] - Fedora-Directory/1.1.2 B2008.248.1443 
starting up

Again, not much help. So I changed the debug level by modifying the 
startup script and adding a -d option to the ns-slapd command. When I 
try to start from the command line, I get this:

Starting dirsrv:
     ldap...[18/Sep/2008:13:57:33 -021800] Fedora-Directory/1.1.2 - 
debug level: accesscontrol (128)
[18/Sep/2008:13:57:33 -021800] - Fedora-Directory/1.1.2 B2008.248.1443 
starting up
Failed to open stats file (/var/run/dirsrv/slapd-ldap.stats) (error 1).
                                                            [FAILED]
*** Warning: 1 instance(s) failed to start

Finally, something helpful! First, why isn't this going to the error 
log? Second, and most importantly, what do I do about it?

I don't know if it's trying to read the file, write the file, do it as 
root, or as nobody. I turned selinux off just to be sure that isn't 
the issue, but I'd prefer to run with it on.

It's a fresh FC9 install, with all patches applied.

TIA,
Dave

--
Fedora-directory-users mailing list
Fedora-directory-users@redhat.com
https://www.redhat.com/mailman/listinfo/fedora-directory-users