tftp service became off after bascul

[wolfgar]

Hi guys,

I faced an issue of CPU switching over on my client PCX. The information of my client is below.

CS type : Appliance server (IBM x306m)
Call server duplication : ip redundancy (both cpus are in same subnet).
Software release : R9.0 h1.301.42a
Shelf type : Common hardware
Phones : IPT4038
Applications : Omnivista4760 (R5.1), CCS (R9.0)
Configuration of netadmin :
<CPUA>
Ethernet interface setup
========================

Netmask : 255.255.255.128

==============================================================================
| Machine type | Local interface | Name | Address |
==============================================================================
| local | Ethernet | axtl-l-hq-001 | 10.3.227.134 |
| local main | Ethernet | axtl-l-hq-002 | 10.3.227.135 |
| router | Ethernet | axa_router1 | 10.3.227.254 |
==============================================================================


<return> to continue.

CPU redundancy setup
====================

Netmask : 255.255.255.128

==============================================================================
| Machine type | Local interface | Name | Address |
==============================================================================
| local | Ethernet | axtl-l-hq-003 | 10.3.227.136 |
| local main | Ethernet | axtl-l-hq-002 | 10.3.227.135 |
==============================================================================

<CPUB>
Ethernet interface setup
========================

Netmask : 255.255.255.128

==============================================================================
| Machine type | Local interface | Name | Address |
==============================================================================
| local | Ethernet | axtl-l-hq-003 | 10.3.227.136 |
| local main | Ethernet | axtl-l-hq-002 | 10.3.227.135 |
| router | Ethernet | axa_router1 | 10.3.227.254 |
==============================================================================


<return> to continue.

CPU redundancy setup
====================

Netmask : 255.255.255.128

==============================================================================
| Machine type | Local interface | Name | Address |
==============================================================================
| local | Ethernet | axtl-l-hq-001 | 10.3.227.134 |
| local main | Ethernet | axtl-l-hq-002 | 10.3.227.135 |
==============================================================================



I performed bascul on main cpu (cpua) after installing new ops. All ipts and GDs were maintained
by cpub as main cpu and confirmed it exactly became main by "twin" and "role -b" command.

However, i confirmed some of issues occurred after bascul

[Issue 1] : Main role address seems not working. (GD and IPT exactly maintained by cpub but 4760
and CCS not able to connect CPUB via main role address. Not able to ping to role address as well)

[Issue 2] : TFTP service on cpua as main cpu never turned "on" after it became main. I confirmed
it by checking the status of xinetd and IPT was not able to connect cpub via main role address
once it had been disconnected.

I tried bascul again after cpua was put in service as standby cpu. But the issue was still after cpua
became main cpu again.
I tried bascul several times after that. However, result was same.

Issues had been solved by following procedures finally.

[Issue 1] : go to netadmin menu and update setting of role addressing as current setting. Just chose
"update" and press enter 3 times...

[Issue 2] : I forced tftp service "on" by command "chkconfig tftp on" and restart xinetd service.

I restored their database in our lab and setup 2 CS as main and standby. However, the issue wasn't
reproduced even if I tried to follow steps what I've done in my client PCX.

Does anyone faced similar issue or any good idea to find the case of issue?

Please advise.