Can anyone tell me why my chained suffix might be hanging after 5 abandoned connections on DSEE 6.2 patch 4? For example, if I use ldapsearch to perform a subtree search against the chained suffix with a filter of objectclass=*, then kill the command with ctrl-c, the 6th attempt to do this will fail. That is, the 6th ldapsearch command hangs with no results returned. If I wait out the nsconnectionlife period (5 minutes), I can run the command again.
I have two servers listed in the nsfarmserverurl attribute for the chain. These are the other specifically configured chain attributes:
nsbindconnectionslimit: 50
nsbindtimeout: 5
nsoperationconnectionslimit: 20
nsconnectionlife: 300
nsslapd-sizelimit: 4000000
The default chain config (dn: cn=default instance config,cn=chaining database,cn=plugins,cn=config) lists the following attribute:
nsabandonedsearchcheckinterval: 2
According to the manual, that attribute defines the number of seconds the server should wait before checking to see if a search has been abandoned. Do I need to specify that attribute explicitly on my chain? Is it perhaps rendered invalid by some other configuration attribute above?
On the servers that I am chaining requests to, I see the following error in the logs:
[21/Jul/2010:13:15:05 +0100] conn=323 op=8 msgId=9 - abandon_core_pb - check_request_controls failed: 12 (Unavailable critical extension) (op=Abandon)
Thanks,
R.
Edited by: lonwt on Jul 21, 2010 5:27 AM
Edited by: lonwt on Jul 21, 2010 5:28 AM
Edited by: lonwt on Jul 21, 2010 7:27 AM