TruCluster
cancel
Showing results for 
Search instead for 
Did you mean: 

5.1B - LAN Interconnect - ics_socket_event: error 60 on channel 0

Gary Hansford
Frequent Advisor

5.1B - LAN Interconnect - ics_socket_event: error 60 on channel 0

I'm running a 2 node Tru64 5.1B ES45 cluster. Interconnect is via GB LAN cards (running on a 100MB fabric).

Every so often the cluster will partition and one node will "yield" to foreign power (tends to be the same one at mo).

Looking at the last /var/adm/messages following appear (bcm0 and bcm1 are netrain'd network, bcm2 is cluster interconnect): -

Sep 12 15:04:04 ukfisch1 vmunix: arp: illegal IP address 255.255.255.255 is used by hardware address 00-80-64-25-54-85!

Sep 12 15:35:53 ukfisch1 vmunix: bcm0: Link up via auto-negotiation (100 Mbps, full duplex)

Sep 12 15:35:53 ukfisch1 vmunix: bcm1: Link up via auto-negotiation (100 Mbps, full duplex)

Sep 12 15:35:53 ukfisch1 vmunix: bcm2: Link up via auto-negotiation (100 Mbps, full duplex)

Sep 12 15:35:53 ukfisch1 vmunix: WARNING: ics_socket_event: error 60 on channel 0, assume node 2 is down

As you can see there was no link down message for any of these ports ! I'm wondering if this has something to do with the cluster_rebuild_delay being 60 (seconds) in the /etc/sysconfigtab ?

Something very weird going on, any suggestions much appreciated...
3 REPLIES
Ralf Puchner
Honored Contributor

Re: 5.1B - LAN Interconnect - ics_socket_event: error 60 on channel 0

could you please post the configuration (e.g. sys_check) within sysconfigtab, rc.config?
How is the gigabit connected (via hub, direct)?
What does netstat -i say?
Are patches installed?
Is the switch/hub proper configured?
who is 00-80-64-25-54-85?
Help() { FirstReadManual(urgently); Go_to_it;; }
Gary Hansford
Frequent Advisor

Re: 5.1B - LAN Interconnect - ics_socket_event: error 60 on channel 0

 
Ralf Puchner
Honored Contributor

Re: 5.1B - LAN Interconnect - ics_socket_event: error 60 on channel 0

the following netstat output indicates a problem:

bcm2 1500 10.1.0 member2-icstcp0 590239310 0 645251948 19123 19123

19123 collisions and 19123 output errors. Please check if the adapter and the switch port is set to the same speed or mode (autoneg).

Verify if the collisions and output errors raises if interconnect problems occurs.

Be sure you have installed the following patches:

BU030627_EW01
REVISION: 0
Mandatory Gigabit Ethernet (DEGXA) Device Driver Update
HP Tru64 UNIX/TruCluster Server 5.1B
PREREQUISITE: Tru64 UNIX/TruCluster Server with PK2 (BL22) installed
ERP Kit Name: T64V51BB22-C0019200-19212-E-20030710.tar
Kit Location: ftp://ftp1.support.compaq.com/public/unix/v5.1b/
Help() { FirstReadManual(urgently); Go_to_it;; }