[Date Prev][Date Next] [Chronological] [Thread] [Top]

Re: slapd no response


>> Our openldap 2.2.24 running on Solaris 9 x86 suddently stopped
>> functioning this morning:( .
>> It can still be started up but doesn't give any response to queries:
>> # ps -ef |grep slapd
>> root 648 1 0 11:18:33 ? 0:00
>> /usr/local/libexec/slapd -h ldap:/// ldaps:///
>> # ldapsearch -x "uid=guest,ou=people,dc=example,dc=com" -wsecret
>> # "uid=tech"
>> ldap_bind: Can't contact LDAP server (-1)
>> Only kill -9 can stop the process. kill -INT or kill -TERM doesn't
>> work.
> Try turning on more logging and using -d on ldapsearch.
> Can you show the full command you use to start slapd? Use /usr/ucb/ps
> -auxwwww Is it showing up on netstat?
> Do you have other system problems? (full disks, etc?)

Since you did a kill -9, you've corrupted the BDB environment, which is why 
it won't restart correctly.

OL 2.2 does not have the automatic BDB database recovery that OpenLDAP 2.3 
has, which is a good reason to upgrade.

You need to find the "db_recover" tool for your specific version of the BDB 
backend. After stopping slapd, cd to your database directory, and run the 
db_recover command. Then start slapd.


  Thanks Quanah. I had to use kill -9 because slapd hadn't responsed to any other signals. In addition,  as I mentioned in my first message, I even tried to recompile the package and ran "make test" (without installing it).  The test failed to start slapd. Does possible corruption of my BDB environment affect the test suite? 
  I have a tape backup of the directory that contains the data files, do you think a simple restore will work? Also I have a LDIF file that contains the database dumped by slapcat. It is helpful to restore the database environment? We are also running a slave server which appears still functional (the slurpd was stopped due to the failure of the master). Can I use it to restore the master server?

Talk is cheap. Use Yahoo! Messenger to make PC-to-Phone calls.  Great rates starting at 1¢/min.