On 05/03/2012 05:06 AM, Roy McMorran wrote:
On 5/2/12 3:57 PM, Paul Graydon wrote:
Intel e1000 drivers have been a continual source of pain for me under
CentOS5. I ended up rebuilding a box as FreeBSD to get past the
problem as the BSD driver was more stable. Didn't matter which
version of the intel drivers I used, as often as 5x a day the network
card would jam up its TX buffers.
Sorry, hijacking the thread a bit, but how could one determine if this
jam-in-the-buffers issue applies to a given server? I see some
occasional short-duration 'freezing' on some RHEL 5 & 6 machines...
haven't been able to id the cause yet. They are using e1000 - maybe
this could be related? Thanks!
It was clear from the dmesg or /var/log/messages output, where you'd see
a line saying "Detected Tx Unit Hang". To be honest, it was short
enough that I probably wouldn't have noticed it, but heartbeat/corosync
did and would fail over with the boxes then proceeding to have an
argument about whether or not they really did 'go away'.
Paul
_______________________________________________
Tech mailing list
[email protected]
https://lists.lopsa.org/cgi-bin/mailman/listinfo/tech
This list provided by the League of Professional System Administrators
http://lopsa.org/
