On Sat, Aug 13, 2016 at 03:40:28PM +0200, mxb wrote:
> > On 11 aug. 2016, at 21:44, Kurt Mosiejczuk <kurt-open...@se.rit.edu> wrote: > > I've noticed that for 5.9, any VMs (in VMware) using vmx(4), end up putting > > "vmx0: device timeout" into the dmesg a bunch when under network load. > > I switched one of the VMs to vic(4) and the messages stop. Another VM that > > I haven't gotten to upgrading from 5.8 to 5.9 yet doesn't show this in its > > dmesg and it's our reasonably busy web server. > > Anyone else see this? I noticed there were a bunch of changes to vmx(4) > > after 5.8 as part of the network mpsafe work. > > dmesg below. > Hey, > it would be nice to define ???network load???. > I have several VMs running 5.8-stable/5.9-stable/current without seeing this. I was easily able to duplicate it when doing an rsync backup to another server. That's when I looked at my IMAP server and noticed it had the same complaint in the dmesg (and have since switched that to vic(4)). Does that describe it better? --Kurt