On 27 November 2014 at 21:09, Garrett D'Amore <[email protected]> wrote:
> Reset failure looks particularly evil. Could be a driver bug. Could be a > broken hba. Or both. > > Sent from my iPhone > > > On Nov 27, 2014, at 8:43 AM, Keith Wesolowski via smartos-discuss < > [email protected]> wrote: > > > >> On Thu, Nov 27, 2014 at 12:20:35PM +0100, Blake Irvin via > smartos-discuss wrote: > >> > >> There hopefully is a core file on your dump device. You should see if > you can coordinate sending that file to Joyent for analysis. The core file > will contain some sensitive information, so don’t post it publicly. See > `man coreadm` for more on dump devices and coredump settings in SmartOS. > > > > You probably want dumpadm(1m) for this, actually. That, and > > savecore(1m). Crash dumps should already be enabled by the OS setup > > utility; dumpadm should confirm that for you. > > > > As for the original problem, I'm not sure. You clearly are having a > > problem between your SAS HBA and one of the targets; you can try to > > decode those codes using the information in the mpt_sas header files. > > As a result of that problem, the HBA's I/O controller was reset. I do > > not know why or how that induced a panic; hopefully that information can > > be gleaned from the crash dump. If you want to debug it yourself, run > > savecore on it and then mdb <number>, where the number is the > > highest-numbered crash dump present (i.e., vmcore.3 -> mdb 3). > > > > Also, if it's the same target every time this happens, you probably have > > a bad disk or possibly a bad disk bay. Replacing that device may work > > around this by avoiding the trigger for the bug (it's still a bug, > > though). > > A quick update. - Somehow the core files were not saved, the dump device changed, this is fixed and if it happens again we will have something to work from - We swapped the disk in the pool out and brought it back to the lab for testing. The disk is broken somehow and shows up a 0bytes in size, not sure if that points to anything in the driver that might cause a panic. Thanks Len ------------------------------------------- smartos-discuss Archives: https://www.listbox.com/member/archive/184463/=now RSS Feed: https://www.listbox.com/member/archive/rss/184463/25769125-55cfbc00 Modify Your Subscription: https://www.listbox.com/member/?member_id=25769125&id_secret=25769125-7688e9fb Powered by Listbox: http://www.listbox.com
