Hi,
This mail is regarding ACE configuration issue.
Please find some info on the setup & the lists of task that we tried to encounter the issue.
1. We have ACE 4710 configured in Bridge mode. ACE 2 interfaces connected with Core Switch. One interface in VLAn 10 (client vlan) & the other in VLAN 100 (server vlan).
2. Only one L3 VLAN 10 is configured in core switch. IN ACE L2 VLAN 10 & 100 is configured for bridging purpose.
3. VLAN 10: Client VLAN
4. VLAn 100: Server VLAN
5. Both of these VLAN's have same IP addressing.
6. There are 2 Cisco Blade Enclosures (Enc1 & Enc 2) connected to Core Switch (4507) in Trunk mode. From individual Enc there are 2 cables connected with core switch.
7. In every Enc there are 8 ports which all in VLAN 100. In every Enc there are 2 blade switches among which 1 is standby & the other active.
8. Please find the IP's for these blade switches.
ENC1:
Sw1: 10.10.10.13
Sw2: 10.10.10.14
ENC2:
Sw1: 10.10.10.15
Sw2: 10.10.10.16
9. There are 2 APP & 2 Staging servers which are in VLAN 10 where as 2 DB servers are in VLAN 100 (all have same VLAN 10 IP)
Problem Facing:
Intermittently, connection from APP server to DB server takes more time (near about 10 seconds). In Ideal condition. it gets connected instantly using SQLPLUS.
Tasks Carried Out:
A. Enabled "Normalization in client vlan 10 in ACE". After enabling this parameter, the frequency of intermittent login to DB from APP got reduced (it is still enabled).
B. Enabled "TCP-Buffer-Share", "TCP-ACK-DELAY","PM-CON-MSS-SET" but it increased the intermittent access frequency so disabled.
C. Sniffed packets. On some occasion it was found from Sniffer that packets received at Client VLAN & in Server VLAN has different TCP sessions for a "similar instance".
Attached docs as follows:
1. ACE config (only normalization enabled on Client & Server VLAN 10 & 100 respectively).
2. ENC1-SW-1
3. ENC1-SW-2
4. ENC2-SW-1
5. ENC2-SW-2
6. SH_LOG_ACE
7. CAPTURE_CLIENT_VLAN
8. CAPTURE_SERVER_VLAN
9. SOW FOR LAOD BALANCERS
Any help on this will be
Rgds,
Partha
