OpenLdap doesn't answer keep-alive from our load balancer

[Chen Shapira <cshapira@mercury.com>]

Hi Everyone,

 

I'm using Openldap with replication between master and slave, and I'm trying
to configure our load balancer (Cisco 11501) to check that our master ldap
is working, and if it isn't, redirect the traffic to the slave ldap.

 

The problem is that ldap doesn't answer keep-alive messages from our load
balancer, so it always report the machine as not-functioning.

When we tried the same load balancer configuration with Sun's ldap, it
worked without issues.

 

Here is what I see in the open-ldap log:

 

connection_get(6224): got connid=122

connection_read(6224): checking for input on id=122

ber_get_next

ber_get_next on fd 6224 failed errno=10054 (WSAECONNRESET)

connection_read(6224): input error=-2 id=122, closing.

connection_closing: readying conn=122 sd=6224 for close

connection_close: conn=122 sd=6224

connection_get(6224): got connid=123

connection_read(6224): checking for input on id=123

ber_get_next

ber_get_next: tag 0x30 len 12 contents:

ber_get_next

do_bind

ber_get_next on fd 6224 failed errno=10035 (WSAEWOULDBLOCK)

ber_scanf fmt ({imt) ber:

ber_scanf fmt (m}) ber:

>>> dnPrettyNormal: <>

<<< dnPrettyNormal: <>, <>

do_bind: version=2 dn="" method=128

send_ldap_result: conn=123 op=0 p=2

send_ldap_response: msgid=2 tag=97 err=2

ber_flush: 52 bytes to sd 6224

 

Anyone tried to do this before? Any ideas what I got wrong?

Thanks,
Chen Shapira
MERCURY
eOperations - Israel
Ext: 2490 cell: 054497262

 



______________________________________________________________________
This email has been scanned by the MessageLabs Email Security System.
For more information please visit http://www.messagelabs.com/email 
______________________________________________________________________