Hi! This is only one driver, mpt2sas downloaded from LSI site version 22.00 instead 13.0 from Debian package. Because from beggining we think that is was a problems in RAID controller. But this hypothesis turned out as erroneous. On other machines we use mpt2sas in-kernel driver only. No 3rd party drivers.
--- С уважением, Буданов Евгений. Системный администратор Компания «Рестрим» On 4 мая 2018 г., 17:08 +0300, Ben Hutchings <[email protected]>, wrote: > Control: tag -1 moreinfo > > On Fri, 2018-05-04 at 15:35 +0300, Eugene Budanov wrote: > > Package: linux-image-4.9.0-6-amd64 > > Version: 4.9.82-1+deb9u3 > > > > Hi! > > > > Here's a short problem description. > > > > We have some Supermicro servers with the same configuration for all > > machines (hardware, kernels, packages, etc). A month ago, or maybe a > > bit later, all of these machines began crashing into kernel panic. I > > can't find any pattern of failure at all. But it happens very often. > > Some machines may drop into kernel panic a couple times a day! But > > usually machines crash about every 3 to 6 days. All of these machines > > have intensive network and i/o operations. > > > > I saved dmesg log from one of these machines after the crash (see the > > attachment). > > > > As far as I see, every machine probably has problems with mlx4_en or > > GRO. Also I see list_add double add => list_del corruption. Can I do > > anything to get more detailed logs? What additional information do > > you need for better problem diagnostics? > > The WARNING messages show that there are out-of-tree modules (i.e. not > part of the kernel package) loaded. What are those? > > Ben. > > -- > Ben Hutchings > Every program is either trivial or else contains at least one bug >
