[389-users] Cannot start the dirsrv process on Debian Lenny

[Rich Megginson]

Morenisco wrote:
> Hi,
>
> I made the following on a Debian Lenny chroot environment:
>
> 1) Installed all the dependences on the OS for the 389-ds-base-1.2.1 
> package.
> 2) I compiled sucessfully the 389-ds-base-1.2.1 package.
> 3) I configured the service, and in the last step the service didn't 
> start, giving me the following error:
>
> Directory Manager DN [cn=Directory Manager]:
> Password:
> Password (confirm):
> /dev/null: Permission denied
> Server failed to start !!! Please check errors log for problems
>
> 4) Error log:
>
> root at dirserv1:/opt/dirsrv/var/log/dirsrv/slapd-dirserv1# cat errors
>    Fedora-Directory/1.2.1 B2009.152.220
>    dirserv1.cdsl.cl:389 (/opt/dirsrv/etc/dirsrv/slapd-dirserv1)
>
> [04/Jun/2009:04:26:02 +0000] - dblayer_instance_start: pagesize: 4096, 
> pages: 524288, procpages: 7193
> [04/Jun/2009:04:26:02 +0000] - cache autosizing: import cache: 204800k
> [04/Jun/2009:04:26:02 +0000] - li_import_cache_autosize: 50, 
> import_pages: 51200, pagesize: 4096
> [04/Jun/2009:04:26:02 +0000] - WARNING: Import is running with 
> nsslapd-db-private-import-mem on; No other process is allowed to 
> access the database
> [04/Jun/2009:04:26:02 +0000] - dblayer_instance_start: pagesize: 4096, 
> pages: 524288, procpages: 7193
> [04/Jun/2009:04:26:02 +0000] - cache autosizing: import cache: 204800k
> [04/Jun/2009:04:26:02 +0000] - li_import_cache_autosize: 50, 
> import_pages: 51200, pagesize: 4096
> [04/Jun/2009:04:26:02 +0000] - import userRoot: Beginning import job...
> [04/Jun/2009:04:26:02 +0000] - import userRoot: Index buffering 
> enabled with bucket size 100
> [04/Jun/2009:04:26:02 +0000] - import userRoot: Processing file 
> "/tmp/ldifJJTzfX.ldif"
> [04/Jun/2009:04:26:02 +0000] - import userRoot: Finished scanning file 
> "/tmp/ldifJJTzfX.ldif" (9 entries)
> [04/Jun/2009:04:26:03 +0000] - import userRoot: Workers finished; 
> cleaning up...
> [04/Jun/2009:04:26:03 +0000] - import userRoot: Workers cleaned up.
> [04/Jun/2009:04:26:03 +0000] - import userRoot: Cleaning up producer 
> thread...
> [04/Jun/2009:04:26:03 +0000] - import userRoot: Indexing complete.  
> Post-processing...
> [04/Jun/2009:04:26:03 +0000] - import userRoot: Flushing caches...
> [04/Jun/2009:04:26:03 +0000] - import userRoot: Closing files...
> [04/Jun/2009:04:26:03 +0000] - All database threads now stopped
> [04/Jun/2009:04:26:03 +0000] - import userRoot: Import complete.  
> Processed 9 entries in 1 seconds. (9.00 entries/sec)
>
> 5) Trying to start it manually:
>
> root at dirserv1:/opt/dirsrv/etc/rc.d/init.d# ./dirsrv start
> Starting dirsrv:
>    dirserv1.../dev/null: Permission denied
> ^C
>
> 6) Reviewing the permissions over /dev/null
>
> root at dirserv1:~/project-389# ls -l /dev/null
> crw-r--r-- 1 root root 1, 3 2009-05-31 23:38 /dev/null
>
> I changed the permissions:
>
> root at dirserv1:~/project-389# chmod 0666 /dev/null
> root at dirserv1:~/project-389# ls -l /dev/null
> crw-rw-rw- 1 root root 1, 3 2009-05-31 23:38 /dev/null
>
> 7) Trying again:
>
> root at dirserv1:/opt/dirsrv/etc/rc.d/init.d# ./dirsrv start
> Starting dirsrv:
>    dirserv1... FAILED
>  *** Warning: 1 instance(s) failed to start
>
> 8) Seeing the error log file:
>
> root at dirserv1:/opt/dirsrv/var/log/dirsrv/slapd-dirserv1# cat errors
> [04/Jun/2009:04:43:11 +0000] - Fedora-Directory/1.2.1 B2009.152.220 
> starting up
> [04/Jun/2009:04:43:11 +0000] - Failed to create semaphore for stats 
> file (/opt/dirsrv/var/run/dirsrv/slapd-dirserv1.stats). Error 
> 38.(Function not implemented)
>
> Some idea bout this error please?
Yes.  This appears to be a problem with setting up chroot environments - 
I have the same problem using mock on rhel/fedora.  Note that you must 
do the mount /dev/shm in the same chroot session as the one you run the 
server - the mount does not persist between chroot sessions, nor does it 
automatically mount.

# sem_open doesn't work
# gets errno 38 (function not implemented)
# I found some information that says this:
# As sem_open() creates named semaphores, it always tries to share them 
between processes.
# Additionally, to support sharing named semaphores with sem_open()
#      add a line to /etc/fstab to mount /dev/shm as a tmpfs
# have to do the mount /dev/shm in the same chroot session
echo tmpfs /dev/shm tmpfs defaults 0 0 >> /etc/fstab
mount /dev/shm

>
> Thanks.
>

-------------- next part --------------
A non-text attachment was scrubbed...
Name: smime.p7s
Type: application/x-pkcs7-signature
Size: 3258 bytes
Desc: S/MIME Cryptographic Signature
URL: <http://lists.fedoraproject.org/pipermail/389-users/attachments/20090604/c567473a/attachment.bin>