Red Hat Bugzilla – Bug 73157
nscd crashes often and therefore fills up logfiles
Last modified: 2008-05-01 11:38:03 EDT
From Bugzilla Helper:
User-Agent: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.0.1) Gecko/20020830
Description of problem:
We are running Red Hat 7.3 and we have installed also Novell eDirectory 8.5.
Novell eDirectory purpose is to allow users to log in to the GNU/Linux-system
with same useraccounts and passwords that they log in to the MS/Windows-system.
Novell eDirectory is huge and buggy software, in my opinion, or we just got
wrong default settings. Anyway, this eDirectory product will make considerable
amount of DNS lookups to manage their weird replicas and so on.
These lookups are very slow if nscd-program is not running. Novell recommends,
if not mandate, to use this program. Unfortunately we have found out that nscd
is very buggy product. It will crash very often.
In fact in our root's crontab reads:
*/2 * * * * /etc/init.d/nscd stop >/dev/null ; /etc/init.d/nscd start >/dev/null
Otherwise the system will not survive. Running the system without nscd is like
swimming in tar.
This leads to another problem, defunct nscd-processes fills the ps-listing. See
attachments. The running versions of buggy nscd also fills up the
/var/logs/messages file. See
Version-Release number of selected component (if applicable):
Steps to Reproduce:
1. Install Novell eDirectory
2. Have nightmares.
Actual Results: Terrible system performance, huge logs and lot's of defunct
Expected Results: Smooth operation with reasonable system resource consumption.
Feel free to ask, I will provide all information you need. Even remote access to
system might be possible to manage.
Created attachment 74244 [details]
Output of 'ps fax' command.
Created attachment 74245 [details]
Few seconds log of /var/log/messages
The log file is /var/log/messages, not /var/logs/messages. Sorry.
Changed crontab to:
*/2 * * * * /etc/init.d/nscd stop >/dev/null ; killall -9 nscd 2>/dev/null ;
/etc/init.d/nscd start >/dev/null
as a work around.
That's all in one line, but bugzilla does the line warping.
This kludge helps quite lot, no more defunct prosesses and no more huge messages
log. Still nscd stability is a problem. Now I need to figure out how often nscd
should be restarted and how many threads should I start.
Files /var/nds/dib/_ndsdb.ini and /nds/dib/_ndsdb.ini now contains following
setup and no other text:
This helps with system speed, but I don't know if it's the optimal setup.
I installed BIND with caching-nameserver settings few days ago. It helps, but
nscd is still unstable according what I have heard from another administator.
Novell sucks. We removed eDirectory. Problem solved, I hope so.