[Linux-cluster] strange modprobe drops VIP & fails service

[Jeff Liu]

I'm experiencing something similar.  I was wondering if there was some
additional information available regarding the "e1000 + multiple NIC
problem".

Thanks in advance,
Jeff


On Fri, 2006-02-10 at 07:16 -0500, danwest wrote:
> Has anyone every seen something like this?  A modprobe with either a
> "module 0" or a blank "module " is always followed by the interface
> loosing its VIP and thus initiating a service failure.  All non
> essential system services were turned off.  Not sure what is
> calling/causing the strange modprobe.  Logs are clean, no other events
> leading up to this.  I am now wondering if it could be the cluster
> itself?
> 
> Feb  9 13:44:32 node2 modprobe: modprobe: Can't locate module 0
> Feb  9 13:44:32 node2 clusvcmgrd: [18150]: <err> service error: IP
> address 147.107.188.81 missing
> Feb  9 13:44:32 node2 clusvcmgrd: [18150]: <err> service error: 0:
error
> fetching interface information: Device not found
> Feb  9 13:44:32 node2 clusvcmgrd: [18150]: <err> service error: Check
> status failed on IP addresses for mservice1

Sounds like the e1000 + multiple NIC problem.

-- Lon