Multiple instances of the NDSD process are seen running when there is only a single instance

  • 7016589
  • 08-Jun-2015
  • 26-Oct-2018


NetIQ eDirectory 8.8 SP8


When running the following command multiple instances of the NDSD process are seen running:
"ps -eaf | grep ndsd"
root      9648     1  0 18:05 ?        00:00:00 /opt/novell/eDirectory/sbin/ndsd
root    10297     1  1 18:07 ?        00:00:00 /opt/novell/eDirectory/sbin/ndsd
Though this can be seen when running multiple instances of eDirectory on the same server, in this case there is only a single instance.


The following are the cases in which this has been seen.
- If a time out has been set for iManager a new instance of NDSD would get spawned while associating a driver to a NCP server object.  (Bug 1073997)  Workaround is to set the timeout to 0 (default).

- An issue in IDM 3.6.1 where a driver would get stuck in a Starting State.  (Bug 854860)
- Some issues have been reported when calling the ndsd binary rather than using the script.  The recommendation going forward is to always use ndsmanage to start and stop eDirectory.
- If the NDSD PID file is manually removed without killing the associated NDSD process the init process has no way to know if a previous instance is already running.  Therefore NDSD will come up again and access the database.   (Bug 639247)
- On OES we now add startproc for the daemon to ensure only one instance can run.  (Bug 828370)
- If two shell sessions are opened and NDSD is simultaneously started from both there is a window in which the PID file from the first instance is not yet created allowing the second instance to start.  (Bug 828370)
- Duplicate lines in the nds.conf file.  Check this file and remove any duplicate statements.
- Lcache crash would spawn a new instance.  (Bug 475914)
- Having an older PA during an upgrade.  There was an issue when upgrading from eDirectory 8.8 SP5 to 8.8 SP6 that if an older PA was in place multiple NDSD processes were started.