We adopted OVS 1.4.3 few weeks ago into XCP 1.1.
No crashes (related to OVS) detected up to now in very large
installation, and old ugly bug with slowly bloating ovs-vswitchd
gone.
We done that process in very slow manner (all hosts was upgraded
within about 6 weeks after 2 week test run on single host and prior
lab testing for about 4 weeks).
Bugs or non-trivial staff found:
1) new ovs do have new policy for MTU handling. Native ovs from XCP
1.1 when found ethX interface unable to change MTU to jumbo sizes,
fall back to original MTU for bridge itself. New OVS now spamming in
dmesg (or messages?) about this, but did not change.
2) It contains new script ovs-vlan-bug-workaround, which failing on
default xs kernel. And our tests shows that OVS is incompatible with
vanilla (upstream) version of ixgbe driver and VLAN's (we using
vanilla instead instead XCP's one).
3) message
Nov 08 02:00:15|72792|netdev|WARN|failed to get flags for network
device vif75.0: No such device
starts to appear in vswitch's log. Still under research, seems be
some /etc/xensource/scripts/vif lines are executed for already dead
vif (no problem so on).
On 07.11.2012 21:24, Jesse Gross wrote:
On Wed, Nov 7, 2012 at 4:58 AM, <frankpaulharms@xxxxxxxxx> wrote:
I've updated OVS on XCP-1.1 to OVS-1.4.2.
I notice random kernel crashes.
Questions:
- Is that simply a bug of 1.4.2 or is there a version limit
regarding the xcp kernel (e.g. not higher than 1.0.2)?
- Which ovs version is recommended if someone likes to
upgrade?
I would try 1.4.3. ÂThere's a bug fix in there that seems
very likely to be the cause of your problem. ÂYou could also
go to 1.7.1.Â
_______________________________________________
Xen-api mailing list
Xen-api@xxxxxxxxxxxxx
http://lists.xen.org/cgi-bin/mailman/listinfo/xen-api
|
