Date: Mon, 28 Dec 2009 17:27:16 +0100 From: Jan Sieka <jps@semihalf.com> To: mjacob@FreeBSD.org Cc: freebsd-stable@freebsd.org Subject: ips(4) in toaster mode FreeBSD 7.2 Message-ID: <4B38DC64.80200@semihalf.com>
next in thread | raw e-mail | index | archive | help
Hi! I'm writing to you because I've seen that you have recently commited patches to the FreeBSD ips(4) driver and perhaps you can shed some light on a problem I've encountered. Here is a description: Recently on one of our servers (IBM xSeries 345 [8760 M1X] with IBM ServeRAID 5i II (Sarasota) RAID controller) ips driver threw a warning about timed-out command and adapter being in toaster mode. After that kernel paniced - see console message below: === Begin of console message === ips0: WARNING: command timeout. Adapter is in toaster mode, resetting to known state ips: io error, status=0x2000c ipsd0: iobuf error 5 ips0: resetting adaptegr_,v ftsh_idso nmea(y) :tiapksed 0usp1 ft[oW R5I TmE(ionftftsese = 1543241728, length=16384)]error = 5 ips0: syncing config Sleeping thread (tid 100006, pid 15) owns a non-sleepable lock panic: sleeping thread cpuid = 2 === End of console message === Lines 5 to 7 are two kernel messages mixed together. They say something like this: ips0: resetting adapter, this may take up to 5 minutes g_vfs_done(): ipsd0s1f[WRITE(offset=154321728, length=16384)]error = 5 After displaying the above messages system is completely unresponsive. The only solution is to reboot. Messages come from functions located in files: "WARNING: [...]": ips_timeout(): sys/dev/ips/ips.c; "resetting adapter, [...]": ips_morpheus_reinit(): sys/dev/ips/ips.c; "syncing config": ips_clear_adapter(): sys/dev/ips/ips_commands.c; I have found someone reporting similar problem (ips in toaster mode throwing a warning and kernel panic after that) to freebsd-stable list in Nov 2006: http://lists.freebsd.org/pipermail/freebsd-stable/2006-December/031469.html The difference is that our server was almost idle (Christmas time) compared to the situation described in the above thread (heavy disk usage during backups). I've checked controller status with IBM's tools (IBM ServeRAID Manager) and it's OK. /var/log/messages yields nothing that could lead to problem's explanation. Server is now up and running, but the reason for this panic is still unclear. I'd be grateful for hints. Also I'd like to know if there are any new changes to be commited to ips driver in future. If that's the case then I will wait for them before applying recent changes to our system. Some info about the system: # uname -a FreeBSD xxx.xxx.xxx 7.2-RELEASE-p4 FreeBSD 7.2-RELEASE-p4 #0: Thu Oct 22 11:01:23 CEST 2009 xxx@xxx.xxx.xxx:/usr/obj/usr/src/sys/XSERIES345 i386 # pciconf -lcv [...] ips0@pci0:8:2:0: class=0x010400 card=0x02591014 chip=0x01bd1014 rev=0x00 hdr=0x00 vendor = 'Elektronik' device = 'ServeRAID 4/5 Morpheus SCSI RAID Controller' class = mass storage subclass = RAID cap 01[80] = powerspec 2 supports D0 D3 current D0 ServeRAID BIOS/Firmware version: 7.12.02 Kernel config is a GENERIC config without unnecessary drivers and features. I'm able to provide any further information about the system if needed. I also send this e-mail to freebsd-stable with hope that somebody has some ideas on my problem. With regards, Jan Sieka
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?4B38DC64.80200>