what' wrong with the system

[leyearn]

the cluster includes two alpha 4100 servers
and hsz50 disk array
on the two server the file /var/adm/syslog.dated/*date/kern.log
the following messages:

Mar 24 11:26:43 bcptt1 vmunix: alt0: Link down
Mar 24 11:26:44 bcptt1 vmunix: alt0: Link up ReceiveFlowControl
Mar 24 11:39:37 bcptt1 vmunix: alt0: Link down
Mar 24 11:39:37 bcptt1 vmunix: alt0: Link up ReceiveFlowControl
Mar 24 13:23:43 bcptt1 vmunix: alt0: Link down
Mar 24 13:23:43 bcptt1 vmunix: alt0: Link up ReceiveFlowControl
Mar 24 13:24:33 bcptt1 vmunix: alt0: Link down
Mar 24 13:24:34 bcptt1 vmunix: alt0: Link up ReceiveFlowControl
Mar 24 13:31:55 bcptt1 vmunix: alt0: Link down
Mar 24 13:31:55 bcptt1 vmunix: alt0: Link up ReceiveFlowControl

the alt0 is 1000Mb/s ethernet interface

[Venkatesh BL]

hmm...looks like a flaky ethernet interface.

[Mark Poeschl_2]

Looks like maybe the switch / network device that the servers are plugged into has some settings that are incompatible with your NIC. If both servers are doing the same thing, check your network device configuration.

[Arch_Muthiah]

Leyearn,
The link is down means that there is a cabling or hub/switch problem.

Try ping to your interface.


Archunan

[Ivan Ferreira]

I have seen this behaviour with cisco switches, sometimes, can be avoided disabling auto-negotiation. But, auto-negotiation should be enabled on GB Ethernet. Anyway, you may try that. As I don't manage the switches, I cannot play with the configuration.

Lastly this could be a bad cable or switch port.