[Date Prev][Date Next]
[Chronological]
[Thread]
[Top]
slapd <defunct> frequently for a moment (ITS#522)
Full_Name: Saurabh Kothari
Version: 1.2.7
OS: RedHat Linux 6.1
URL: ftp://ftp.openldap.org/incoming/
Submission from: (NULL) (202.142.88.6)
We are using OPENLDAP with Livingstone RADIUS 2.2 for authentication of Dial-Up
users to an ISP setup. Other Netscape API's also poll the LDAP server for users
information such as online user creation, users changing there own password and
online user account information.
As per the former BUGS posted for slapd <defunct> it shows the problem was with
avalibility of space on /var and other outcome was problem with Java cookies
sync. But in our case both are not in dought as /var has 3.4 GB and there is no
use of Java Cookies.
The problem we face is slapd shows <defunct> for a moment and then goes away
after a random period of time which can be 15 seconds to 60 seconds. Then after
another random period of time of 1 to 5 minutes it again shows the same thing.
Due to which the user tring to login at that moment is not allowed and in the
radius log it is says can't communicate to the LDAP server. But then at next
instance it allows the user to login as the <defunct> process goes off. This is
happing so frequently due to which sometimes the users get error of user already
logged in.
The system condition is :
There are already three process of slapd and four process of slurpd running in
the system
-----------------------------------------------------------------------------------
root 5305 0.0 0.2 4536 2732 ? S 11:30 0:00 slapd
root 5306 0.0 0.2 4536 2732 ? S 11:30 0:00 slapd
root 5307 0.0 0.2 4536 2732 ? S 11:30 0:04 slapd
root 5319 0.0 0.0 1656 808 ? S 11:30 0:00 slurpd
root 5320 0.0 0.0 1656 808 ? S 11:30 0:00 slurpd
root 5321 0.0 0.0 1656 808 ? S 11:30 0:00 slurpd
root 5322 0.0 0.0 1656 808 ? S 11:30 0:00 slurpd
----------------------------------------------------------------------------------
and then at the end it show a slapd <defunct>
==================================================================================
root 18803 0.0 0.0 0 0 ? Z 17:15 0:00 [slapd <defunct>
==================================================================================
Sometimes the forth process of slapd comes properly with R status but at that
time the second process of sldap also has a R status instead of S status.
===================================================================================
root 5305 0.0 0.2 4536 2732 ? S 11:30 0:00 slapd
root 5306 0.0 0.2 4536 2732 ? R 11:30 0:00 slapd
root 5307 0.0 0.2 4536 2732 ? S 11:30 0:04 slapd
root 5319 0.0 0.0 1656 808 ? S 11:30 0:00 slurpd
root 5320 0.0 0.0 1656 808 ? S 11:30 0:00 slurpd
root 5321 0.0 0.0 1656 808 ? S 11:30 0:00 slurpd
root 5322 0.0 0.0 1656 808 ? S 11:30 0:00 slurpd
-----------------------------------------------------------------------------------
root 18961 0.0 0.2 4536 2732 ? R 17:19 0:00 slapd
===================================================================================
We have enabled the loglevel in slapd.conf with 256 value to trace the problem
some of the output's are as under:
===================================================================================
May 2 17:39:04 radblr slapd[19878]: conn=1861 op=0 BIND
dn="UID=HELPDEL,OU=PEOP
LE,DC=ZEEACCESS,DC=COM" method=128
May 2 17:39:04 radblr slapd[19878]: conn=1861 op=0 RESULT err=49 tag=97
nentrie
s=0
May 2 17:39:05 radblr slapd[19879]: conn=1861 op=1 BIND
dn="UID=HELPDEL,OU=PEOP
LE,DC=ZEEACCESS,DC=COM" method=128
May 2 17:39:05 radblr slapd[19879]: conn=1861 op=1 RESULT err=49 tag=97
nentrie
s=0
May 2 17:39:05 radblr slapd[19880]: conn=1861 op=2 UNBIND
May 2 17:39:05 radblr slapd[19880]: conn=1861 op=2 fd=28 closed errno=0
-----------------------------------------------------------------------------------
May 2 17:49:31 radblr slapd[20247]: conn=1907 op=0 BIND
dn="UID=IMPULSE,OU=PEOP
LE,DC=ZEEACCESS,DC=COM" method=128
May 2 17:49:31 radblr slapd[20247]: conn=1907 op=0 RESULT err=48 tag=97
nentrie
s=0
May 2 17:49:32 radblr slapd[20248]: conn=1907 op=1 BIND
dn="UID=IMPULSE,OU=PEOP
LE,DC=ZEEACCESS,DC=COM" method=128
May 2 17:49:32 radblr slapd[20248]: conn=1907 op=1 RESULT err=48 tag=97
nentrie
s=0
May 2 17:49:33 radblr slapd[20249]: conn=1907 op=2 UNBIND
May 2 17:49:33 radblr slapd[20249]: conn=1907 op=2 fd=23 closed errno=0
May 2 17:49:59 radblr slapd[20253]: conn=5 op=3641 SRCH
base="DC=ZEEACCESS,DC=C
OM" scope=2 filter="(uid=TELERATE)"
May 2 17:49:59 radblr slapd[20253]: conn=5 op=3641 RESULT err=0 tag=101
nentrie
s=1
===================================================================================
One observation from my end is whenever ladp tries to generate the fourth
process of slapd it goes <defunct> due to which the process ID of the <defunct>
process keeps on continously changing (increamenting).
If more info. or full log is required kindly let me know.
Saurabh Kothari