[prev in list] [next in list] [prev in thread] [next in thread] 

List:       linux-ldap
Subject:    slapd hangs on Redhat 5.0
From:       jpm () Retail-SC ! com (Jan-Piet Mens)
Date:       1998-07-01 8:39:39
[Download RAW message or body]

Hello,

I'm using slapd from ldap-3.3-4.i386.rpm on a Redhat 5.0 system. I have two
ldbm backends, one with only a dozen entries and the main ldbm with 13000
entries.

slapd is visited by all sorts of tools: ldapsearch, a WWW interface, an MTA, etc.
The main backend has 8 indices. Once in a while, slapd stops working. The 3
processes are visible in the process list, but no connections are accepted
(they aren't refused either!). The log file shows

Jun 30 22:21:32 gin slapd[14202]: conn&3 op=0 BIND dn="" method8
Jun 30 22:21:32 gin slapd[14202]: conn&3 op=0 RESULT err=0 tag— nentries=0
Jun 30 22:21:32 gin slapd[14203]: conn&3 op=1 SRCH base="o=Org,c=XX" scope=1 filter="(|(uid=JPM)(alias=JP
M))"
Jun 30 22:21:32 gin slapd[14203]: conn&3 op=1 RESULT err=0 tag1 nentries=1
Jun 30 22:21:32 gin slapd[14205]: conn&3 op=2 SRCH base="o=Org,c=XX" scope=1 filter="(|(uid=JPM)(alias=JP
M))"
Jun 30 22:21:32 gin slapd[14205]: conn&3 op=2 RESULT err=0 tag1 nentries=1
Jun 30 22:21:32 gin slapd[1407]: pthread_create failed
Jun 30 22:44:12 gin slapd[1407]: conn&4 fd  connection from unknown (127.0.0.1)
Jun 30 22:44:12 gin slapd[1407]: pthread_create failed
Jun 30 22:49:12 gin slapd[1407]: conn&4 op=-1 fd  closed errno=0
Jun 30 22:50:32 gin slapd[1407]: conn&5 fd  connection from unknown (127.0.0.1)
Jun 30 22:50:32 gin slapd[1407]: pthread_create failed
Jun 30 22:55:32 gin slapd[1407]: conn&5 op=-1 fd  closed errno=0
Jun 30 22:55:33 gin slapd[1407]: conn&6 fd  connection from unknown (127.0.0.1)
...

I have to stop and restart slapd to get it back to normal.
What I don't understand, is that I have an identical backup server; same hardware,
same memory, same Redhat 5.0, same glibc, same slapd, same backends, same database. This
system does not show these symtoms. I've beaten the hell out of the second server,
retrieving thousands of entries without it hanging up.

Any ideas ?
I would certainly appreciate any help you can give me.

Regards,
	-JP

[prev in list] [next in list] [prev in thread] [next in thread]