This message was deleted.
# citrix-netscaler
s
This message was deleted.
e
We balance LDAP successfully with lb vs. leastconnection. No persistence. All parameters default except ciphers for 636. 636 is running on ssl_tcp. That’s it.
j
Very good, thanks a mill!
I have reports of timeout and slowness but using the LDAP VIP. I suspect surge queue, SP is disabled, it is strange. I can see the surge queue in NS graph spiking a bit and then going down. The load is not that heavy but I think NS is probably reaching some limits but I can't really see something proving this.
e
Difficult to tell you what’s the problem without deep analysis. Could be that you get some requests queueing up anyhow. I have also heard for example of strange behavior of ADC when running on Hyper-V without min 4GB Memory per Core. Would suggest to run min 2 cores with 8GB Memory at least for VPX.
j
Indeed thanks, I was looking for some hints on what I could check to see if I don't hit the max nb established clients / servers connections. That's the only thing higher. It's a MPX 7500, CPU/Memory it's all good.
e
Very old box. I would check what this box can handle in SSL TPS when you run LDAPS as SSL or SSL_TCP vServer. could reach out some limitation if you have high number of handshakes and using strong ciphers.
I will check if I find the nsconmsg command I used in the past to check for queuing
nsconmsg -K /var/nslog/newnslog -g ssl_cur_sslInfo_nsCardInQCount -d current Was using this to identify ssl queueing
j
Thanks yes I was using it, I can't see anything wrong and the issue (time out, very slow requests initialization(prob in the queue)) now for 2 different apps started when using this VIP. It is LDAP 389, not SSL. I'll keep digging.