Instance Number 1, installed in /export/home/omni, has not completed configureNodes

xemacs /etc/swinstall/db/1/instance_state

Change variant without running configurePlatform

DFedit portConf.$SHM
DFedit portConf.GENERIC

Know whether a logical node is started / stopped

    SYSlogicalNodeList_t actv_node_list;
    memset(&actv_node_list, 0, sizeof(SYSlogicalNodeList_t));
    if (RETURNerror != SYSgetNodeNames(&actv_node_list, TRUE))
    {
        for (int j = 0; j < actv_node_list.count; j++)
        {
            if (!strcmp(actv_node_list.node[j].name, logicalNode))
            {
                /* Node is started. */
                ...
                break;
            }
        }
    }

Managers by type of logical node

Error message “command not valid on this logical node” is issued by NM (see nmhandlers.c) after SYSgetNodeConfInfo(), which reads the list of managers from nodeStart.myNode.$SHM

Based on api_main.c.

   C71 (         C7)                MTP3           SCCP     TUP*    IUP*   ISUP*   TCAP 
 C7MSG (   C7M3UASG)     M3UA SG    MTP3           SCCP*    TUP*    IUP*   ISUP*   TCAP*
 C7MAS (     C7M3UA)     M3UA AS                   SCCP     TUP*    IUP*   ISUP*   TCAP 
 C7SSG (    C7SUASG)                MTP3  SUA SG   SCCP     TUP*    IUP*   ISUP*   TCAP 
 C7SAS (      C7SUA)                      SUA AS                                   TCAP 
C7MSSG (C7M3UASUASG)     M3UA SG    MTP3  SUA SG   SCCP     TUP*    IUP*   ISUP*   TCAP 
C7MSAS (  C7M3UASUA)     M3UA AS          SUA AS            TUP*    IUP*   ISUP*   TCAP 

Notes:

• This manager* is optional
• SCCP manager => TCAP manager
• TUP and IUP are antinomic
• Hybrid stack options, TUP, IUP, ISUP options are proposed only if packages installed (see options.GENERIC)

Non-applicable parameters

Q: To the file $OMNI_HOME/conf/omni_conf_info on both CEs I have to add

MOPC_GR=YES

MOPC is only supported for protocols A7,C7, AC7, CA7, and CH7. Does this entry harm for other protocols like M3UA?

A: Prior to SP4 the presence of this parameter in the omni_conf_info file when running M3UA would have resulted in a “non-provisioned” error in the diag file. From SP4 onwards you will no longer see the error.

See configureNodes flow

export VERBOSE=1
configureNodes ...

configureNodes: Please run configurePlatform

1. Ensure /etc/swinstall/db/1/instance_state has all its statuses set to xxx_COMPLETE
2. Ensure $OMNI_HOME/conf/configurePLatform.Conf is older than $OMNI_HOME/conf/ce.conf.GENERIC

Monta Vista queues

Hong Ye was having problems with DFdaemon on his MV boxes in a M3UA configuration.
( M3UAmgr passes 10meg files across the CEs using DF… )

Eric noticed on Hong Ye’s
Monta Vista boxes that ”/sbin/ifconfig -a” cmd shows that the “txqueuelen”
parameter of each “ethx” interfaces was set to 100 whereas it is set to 1000 on other Linux machines.

Small values for txqueuelen are OK for slow device but not for Ethernet links (see ifconfig man page).

Using ifconfig cmd, Eric set txqueuelen=1000 on all “ethx” interfaces of Wilson and Roosevelt.
Everything is OK now.

It looks like txqueuelen=100 is the default for Monta Vista machines.
txqueuelen=1000 should be set by default on ALL Linux machines.

SS7 SLK stay failed

DFcat multiconf.cnf

and check that we have:

CLK_SOURCE TDMA:S,TDMB:M

and not

CLK_SOURCE TDMA:S,TDMB:N

CREATE-M2PA-PORT error: Invalid local

C71:lili:698 > CREATE-M2PA-PORT:NAME=PORT2,LOCAL="lino",REMOTE="xenon",SCTP_PORT=1118,SCTP_SERVER_PORT=3672,TYPE=SERVER;
M  DENY
  IDNV
  PORT_DAEMON: Invalid local

Then, in /etc/hosts, replace

127.0.0.1       localhost.localdomain   localhost       lino

with

127.0.0.1       localhost.localdomain   localhost

and

C71:lili:698 > CREATE-M2PA-PORT:NAME=PORT2,LOCAL="lino",REMOTE="xenon",SCTP_PORT=1118,SCTP_SERVER_PORT=3672,TYPE=SERVER;
M  COMPLETED

25-Jun-2009 23:02:08 spurs.MOM Error 00010 00005 Queuing not allowed O:spurs.DR Q:0×83f0000 T:0×301b, suppressed 0

This occured while stopping, then re-starting the STBY CE. Following this, the STBY CE was dying.

Hello Hong,

eth3 interfaces on both knicks and spurs are not configured correctly.

We can see that IP packets were dropped on eth3:

knicks

eth3     Link encap:Ethernet  HWaddr 00:0E:0C:EC:AC:89  
         inet addr:10.2.89.4  Bcast:10.2.89.255  Mask:255.255.255.0
         inet6 addr: fe80::20e:cff:feec:ac89/64 Scope:Link
         UP BROADCAST RUNNING MULTICAST  MTU:1500  Metric:1
         RX packets:1719575 errors:0 dropped:2086 overruns:0 frame:0
         TX packets:1156494 errors:365 dropped:0 overruns:0 carrier:365
         collisions:31041 txqueuelen:100
         RX bytes:1246952841 (1.1 GiB)  TX bytes:474268341 (452.2 MiB)
         Base address:0x4000 Memory:bc520000-bc540000

spurs

eth3     Link encap:Ethernet  HWaddr 00:0E:0C:EC:AD:99  
         inet addr:10.2.89.3  Bcast:10.2.89.255  Mask:255.255.255.0
         inet6 addr: fe80::20e:cff:feec:ad99/64 Scope:Link
         UP BROADCAST RUNNING MULTICAST  MTU:1500  Metric:1
         RX packets:1155753 errors:0 dropped:68 overruns:0 frame:0
         TX packets:1721684 errors:19 dropped:0 overruns:0 carrier:19
         collisions:31041 txqueuelen:100 
         RX bytes:474054722 (452.0 MiB)  TX bytes:1250025688 (1.1 GiB)
         Base address:0x4000 Memory:bc520000-bc540000

I ran ifconfig eth3 txqueuelen 1000 on both.

But the main issue here is that eth3 is 100 Mbps Half Duplex where eth4 is 1000 Mbps Full Duplex:

> grep "Half Duplex" /var/log/messages

Jul  6 14:18:54 spurs kernel: e1000: eth3: e1000_watchdog_task: NIC Link is Up 100 Mbps Half Duplex
Jul  6 14:34:44 knicks kernel: e1000: eth3: e1000_watchdog_task: NIC Link is Up 100 Mbps Half Duplex

This must be fixed.

Eric.