Dale Shaw wrote:
I'm sorry if this is a FAQ. I searched but couldn't find a direct
match yada yada..
I have a vanilla 6.0-RELEASE system running a bunch of network
management type tools like RANCID, nfcapd, cacti and so on.
After a few days of normal operation, the system (locked away in a
data centre) falls off the network. Can't SSH to it, can't ping it. No
ARP -- gone! I have no OOB access to this machine (it's a test
Strangely, when I drive out and visit the machine (a HP DL320) in
person and press enter a couple of times on the keyboard, it springs
back to life like nothing ever happened. I literally see (for example)
log entries dated "Feb 07" immediately followed by entries dated "Feb
28". Some processes lose the plot and need to be restarted, but others
just continue on their merry way.
In my searching I have found a couple of references to dodgy keyboard
drivers, problems with systems on KVM switches (like this one is), and
power management issues.
Any clues? Unfortunately tracking -STABLE is not really an option for
me on this box, but I'm more than happy to update and build a new
kernel on a once-off basis if someone says it's a known bug/problem
sorted out in a post-6.0-RELEASE fix.
I have a vague memory of something like this but cannot point to a
specific commit that resolved it.
email@example.com mailing list
To unsubscribe, send any mail to "[EMAIL PROTECTED]"