This discussion has been locked.
You can no longer post new replies to this discussion. If you have a question you can start a new discussion

OES 2015sp1 ndsd very slow to start after updates and reboot

OES2015sp1/SLES11sp4 on VMware 6.5
Online updates installed last night - last updated in Nov 2017.
After update install finished - server was rebooted - as required.
After reboot - very slow starting
Stalled for a long time at Executing customized settings after starting the NetIQ eDirectory server…
Then shows:
Starting Novell Storage Services (NSS)
nss: ERROR: required eDirectory(ndsd) is not running. Exiting.
………. ABORTING /sbin/rcnovell-nss ……….

I suspect a time issue.
When server restarts messages log shows UTC time which then reverts to Eastern (-04:00) soon after.
This setup is configured the same as other sites that are running OES11sp2/SLES11sp3 on VMware 5.5 with respect to all time settings - that I can find.

Any thoughts?
  • There are quite a few things which can go wrong as for time stuff, most them get evident when a box gets rebootet after a shift to or from DST. But before chasing in this direction: did the server certificates expire by chance?
  • Server certificate is good.
    Additional info:
    After 4 hours elapsed novell-named started (when it caught up to the time showing in the log just after restarting)
    I manually started novell-nss in the morning without any problem.
  • Things to check:
    - make sure that the hardware clock(s) of the ESX host(s) is / are set to UTC
    - make sure the virtual bios clock of the VM is the to UTC
    - in /etc/sysconfig/clock check for HWCLOCK="-u", SYSTOHC="yes" and the proper timezone
    - in /etc/sysconfig/ntp check for NTPD_FORCE_SYNC_ON_STARTUP="yes" and NTPD_FORCE_SYNC_HWCLOCK_ON_STARTUP="yes"
    (of course assuming a reliable ntp source)

    maybe you don't have a chance to bounce the entire host right now, but at least you can
    - make sure ntp is configured properly and syncing (if not -> configure it to do so)
    - stop ndsd
    - stop ntpd
    - make the changes to the sysconfig files
    - down the VM
    - configure the VM to start into the virtual bios on the next startup
    - start the VM, make sure bios time is set to UTC, save the changes
    - let the VM boot
  • Thanks for the info.

    I have never set this before
    " in /etc/sysconfig/ntp check for NTPD_FORCE_SYNC_ON_STARTUP="yes" and NTPD_FORCE_SYNC_HWCLOCK_ON_STARTUP="yes" "
    but I will give it a try.

    Everything else, except checking with the VM BIOS, is OK.
    Time is showing correctly everywhere I look.

    It may be a few days before I get a chance to reboot the host server or the VM.
  • All these items (VM BIOS is vital, too) will help to avoid that eDir initialization falls into a gap where the system time is simply wrong. The latter would (depending on your timezone (really)) cause stuff such as VERY delayed init of ndsd, ndsd skipping SLP registrations on every 2nd. cycle and little purple monster creeping around in the racks.
  • Thank you Mathias for pointing me in the right direction.
    I had a chance to reboot the host and the guest last night.
    The guest VM BIOS time was 4 hours fast. The difference between UTC and current local time.
    Somewhere in the initial setup the VM BIOS must have got local time instead of UTC.
    OES server boots normally now and all services start as required.
  • Good to hear, thanks for reporting back.