Hi there,
Is this a bug? We’ve got a 10.6.4 Server OD Master setup with round about 800 users, 12GB RAM that sometimes starts to get unresponsive. It’s slapd that suddenly refuses to respond. CPU load rises to around 10-15% which is high for Intel Xserves. When this happens, our mail service can’t authenticate anymore (we use communigate with an external authentication helper script that is designed to restart every 30 seconds. If we restart slapd, all is fine again for some days up to a week or so.
DNS is all fine, Kerberos and password server ok, changeip -checkhostname results also in an ok. The server was setup from scratch with imported groups and users from a 10.5.8 od master. I’ve already increased the the amount of returned values of the ldap to 25.00 that didn’t solve the issue because we first thought the external helper script of our mail server causes the problem, but it’s in fact slapd that hangs. What could trigger this behavior with a brand new Xserver with 12GB RAM, with all services configured perfectly?
Excerpt of the error log:
request: socketpair failed 24 (Too many open files)
Aug 27 11:50:51: — last message repeated 499 times —
Aug 27 11:50:51 authority slapd[92258]: *** process 92258 exceeded 500 log message per second limit – remaining messages this second discarded ***
Aug 27 11:50:52 authority slapd[92258]: dnssd_clientstub deliver_request: socketpair failed 24 (Too many open files)
Aug 27 11:50:52: — last message repeated 499 times —
Aug 27 11:50:52 authority slapd[92258]: *** process 92258 exceeded 500 log message per second limit – remaining messages this second discarded ***
Aug 27 11:50:53 authority slapd[92258]: dnssd_clientstub deliver_request: socketpair failed 24 (Too many open files)
Aug 27 11:50:53: — last message repeated 499 times —
Aug 27 11:50:53 authority slapd[92258]: *** process 92258 exceeded 500 log message per second limit – remaining messages this second discarded ***
Any idea what could be the reason and how to fix that?
Alex
Comments are closed