Date: Fri, 7 Oct 2011 13:46:36 -0700 From: Jason Wolfe <nitroboost@gmail.com> To: Arnaud Lacombe <lacombar@gmail.com> Cc: freebsd-net@freebsd.org Subject: Re: Intel 82574L interface wedging on em 7.1.9/7.2.3 when MSIX enabled Message-ID: <CAAAm0r1ed1U=LzYhHfCfqr2N9MmC6OR1qSk-wZVCkiaPQ0uvaw@mail.gmail.com> In-Reply-To: <CACqU3MVUsO4eRS8rZannCO%2BKRf0T55aaYSXrkavi07LMofwreQ@mail.gmail.com> References: <CAAAm0r0RXEJo4UiKS=Ui0e5OQTg6sg-xcYf3mYB5%2Bvk8i8557w@mail.gmail.com> <CAFOYbc=6J%2B43sqMEO3gYGQXdOaOw2V4sV=xdgr3feKcc3qPM3A@mail.gmail.com> <CAAAm0r2MoFrGE_ubB_CX8jrV9cxobR-rsrZbRiRqBiLO_-Py4w@mail.gmail.com> <CACqU3MVUsO4eRS8rZannCO%2BKRf0T55aaYSXrkavi07LMofwreQ@mail.gmail.com>
next in thread | previous in thread | raw e-mail | index | archive | help
On Fri, Oct 7, 2011 at 12:24 PM, Arnaud Lacombe <lacombar@gmail.com> wrote: > Hi, > > On Fri, Oct 7, 2011 at 2:57 PM, Jason Wolfe <nitroboost@gmail.com> wrote: > > Jack, > > > > Entirely possible there are multiple moving pieces here, the only bit I > know > > for certain is it's related to the different operation when running with > MSI > > vs MSI-X. Here is also my loader.conf for reference. I'm currently > running > > the modular congestion control stuff with cubic in use, but these issues > > predate those changes also. Just to give you a scope of it though, it was > > somewhat 'rare' for them to wedge. Out of a pool of ~2000 servers running > > with the 82574L doing ~800Mb/s average, there were ~220 reports in a > week. > > So with some fuzzy math to put it in the same terms you were talking in, > a > > server in particular would hang about once every 9 weeks. > > > Just a two questions out of my mind: > > Are the failing server evenly distributed, or always the same are failing ? > > Did you collect the uptime and the kernel msgbuf of the server when > the issue triggered ? > > Thanks, > - Arnaud > Arnaud, The failures were pretty random, though there were a handful of servers that did fail a couple times. It didn't seem attributable to a certain batch or physical location. The uptime was not collected, but most were in the ballpark of 30-90 days. I was tailing /var/log/messages, but didn't save kern.msgbuf no. I've added both of these to the collections and pulled a couple that did fail more than once and will be re enabling MSI-X on them later today. Jason
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?CAAAm0r1ed1U=LzYhHfCfqr2N9MmC6OR1qSk-wZVCkiaPQ0uvaw>