[Search for users] [Overall Top Noters] [List of all Conferences] [Download this site]

Conference smurf::ase

Title:ase
Moderator:SMURF::GROSSO
Created:Thu Jul 29 1993
Last Modified:Fri Jun 06 1997
Last Successful Update:Fri Jun 06 1997
Number of topics:2114
Total number of notes:7347

1851.0. "ASE V1.4 and Monitored interface ?" by NETRIX::"[email protected]" (Albertino) Thu Jan 30 1997 12:17

Hi,


A customer enconters a  problem with ASE v1.4 .

If he carry off the network cable on a monitored interface on a system A which
own 3 services the 3 services doesn't relocate on system B .

The configuration follows:

There is two systems : NODE A -> hostname :emeraude ;Digital Unix v4.0A ASE
1.4  
		       NODE B  -> hostname diamant ;Digital unix v4.0A  ASE 1.4 

There is two networks and 3 services on each node.

primary network 130.10
secondary network 130.11

#cat /etc/hosts
~
~
130.10.31.2 emeraude
130.10.31.1 diamant
130.11.31.2 emeraude0
130.11.31.1 diamant0
~
~
#cat /etc/rc.config (on  node diamant)
|
net_dev_0=tu1 --->primary network
net_dev_1=tu0 --->seconadry network
ifconfig_0=130.10.31.1 netmask 255.255.0.0
ifconfig_1=130.11.31.1 netmask 255.255.0.0

#cat /etc/rc.config (on node emeraude)
net_dev_0=tu1
net_dev_1=tu0
ifconfig_0=130.10.31.2 netmask 255.255.0.0
ifconfig_1=130.11.31.2 netmask 255.255.0.0

#cat /etc/routes
-host diamant diamant
-host diamant0 diamant0
-host emeraude emeraude
-host emeraude0 emeraude0

With asemgr :

name member	name interface	member net	monitor

emeraude	emeraude	primary		yes
emeraude0	emeraude0	backup		no

diamant		diamant		primary		yes
diamant0	diamant0	backup		no

(The ip adresses of all services are like 130.10.xx.xx (network 130.10))

ASEROUTING is not set

If we carry off the network cable on primary monitored interface on , per 
exemple on system emeraude , the 3 services on emeraude doesn't relocate on
diamant.

the last entry in daemon.log when it occurs is 

HSM_PATH_STATUS:130.10.31.1: DOWN 130.11.31.1 is UP


 I see the page 1-9 of "the Managing an available server
Environment" it is said that:

>you can monitor specific network interface and take specific actions
(such as relocating services)when a particular interface fails.
~
~if a monitored network interface fails trucluster runs the error Alert script
which invokes /var/ase/lib/ni_status_awk script....
the default script causes trucluster to stop all the services running on that
member system and start them on another system .....



Also what i missed ? Is it normal ?

Thanks for help
ALbertino

[Posted by WWW Notes gateway]
T.RTitleUserPersonal
Name		DateLines
1851.1Only one network monitored ? for sure ?BACHUS::DEVOSManu Devos DEC/SI Brussels 856-7539Fri Jan 31 1997 04:0918
    Hi Tino,
    
    >> HSM_PATH_STATUS:130.10.31.1: DOWN 130.11.31.1 is UP     
    
    You say the above is the last message in the daemon.log..., but we need
    to know the previous messages, particularly one like:
    
            HSM_NI_STATUS:130.10.31.1:XXXX:130.11.31.1:XXXX 
    
    
    I am curious to see if it reports ONE IP address or two. According to
    the network setup you showned us, it should only contain one IP address
    and then you are rigth, the services should failover, but if it is
    showing two IP addresses (one UP and one DOWN) then not all interfaces
    are down and your services should continue (and the network setup you
    showned  is incorrect)
    
    Regards, Manu.
1851.2trace daemon.logNETRIX::"[email protected]"AlbertinoFri Jan 31 1997 09:4831
I don't see anything about HSM_NI_STATUS the trace follows :

Jan 29 21:47:25 emeraude ASE:local HSM Warning:Can't ping diamant over the 
network 

Jan 29 21:47:26 emeraude ASE:local HSM notice:/var/ase/sbin/ase_run_sh:change 
host 130.10.31.1: gateway 130.11.31.1

Jan 29 21:47:26 emeraude ASE:local HSM Notice:/var/ase/sbin/ase_run_sh:change
host 130.11.31.1:gateway 130.11.31.1

Jan 29 21:47:26:emeraude ASE:local HSM ***ALERT:
HSM_PATH_STATUS:130.10.31.1:DOWN:130.11.31.1:UP

Jan 29 21:51:01 emeraude ASE:local HSM Notice:Able to ping diamant over the
network

Jan 29 21:51:01 emeraude ASE:local HSM Notice:/var/ase/sbin/ase_run_sh:change 
host 130.10.31.1:gateway 130.10.31.1
Jan 29 21:51:01 emeraude ASE:local HSM Notice:/var/ase/ase_run_sh:change host
130.11.31.1:gateway 130.11.31.1

Jan 29 21:51:01 emeraude ASE:local HSM***ALERT:
HSM_PATH_STATUS:130.10.31.1:UP:130.11.31.1:UP


That's all 

Albertino
[Posted by WWW Notes gateway]
1851.3BACHUS::DEVOSManu Devos DEC/SI Brussels 856-7539Tue Feb 04 1997 04:1711
So Tino,

You should review the network setup. It appears that 

>>> diamant		diamant		primary		yes

this line is probably set to no at the end of the line.

Can you check the monitored network?

Manu.
1851.4I'm sure ..NETRIX::"[email protected]"AlbertinoTue Feb 04 1997 12:2514
Hello ,

I phone to my customer , ask him to verify the network configuration and 
the configuration is like as i say in .0 .

emeraude  emeraude  primary  Yes
emeraude  emeraude0 backup    no

diamant  diamant  primary  yes
diamant  diamant0  backup  no

Albertino
[Posted by WWW Notes gateway]
1851.5I'm sureNETRIX::"[email protected]"AlbertinoThu Feb 06 1997 05:1915
Hello,


 I phone to my customer and we verify the  ASE Network Configuration together
.
 The configuration is like i said in .0.


(We try in our lab with a memory channel configuration and with the tulip 
interface on a AS2100 and we doesn't reproduce the problem and we saw the 
 ASE: local HSM ***ALERT: HSM_NI_STATUS: message in daemon.log)  

 Albertino
[Posted by WWW Notes gateway]
1851.6Without HSM_NI_STATUS, no chance...BACHUS::DEVOSManu Devos DEC/SI Brussels 856-7539Fri Feb 07 1997 08:577
    Hello Tino,
    
    In this case, ask your customer to repeat the test, but you MUST see a
    HSM_NI_STATUS message otherwise the system will not react as you
    indicated in .0
    
    Regards, Manu.