Hi Michele, On Wed, 2014-11-19 at 22:48 +0100, mpe wrote: > Hi all, > One of our systems is seeing a kernel BUG with a fibre channel target > system. The system is an xeon 56xx with a qle2462 and the initiator is > vmware 5.x. > The system is ubuntu 14 LTS and we've configured FC with targetcli with > multipath enabled. > The starting configuration it's that we have both FC channel connected, > vmware sees them and without problems I can use the exported LUN for > make my bonnie++ tests. bonnie++ are executed inside a virtual machine > that has it's hard disk inside the LUN exported. > For our multipath tests we detach one of FC optical patch, the system > (vmware) waits about 20 seconds and the after bonnie++ writes continue. > > The bug/crash is repeatable and it happens every time I reconnect FC > optical patch: after about 30/40 sec, FC is renegotiated and after other > 30/40 sec, the system becomes unresponsive and I receive a lot of kernel > log msg like: > > Performing ISP error recovery - ha=ffff880466e5c000. > qla2xxx [0000:10:00.0]-505f:9: Link is operational (4 Gbps) > TARGET_CORE[qla2xxx]: Detected NON_EXISTENT_LUN Access for 0x00000001 > Error entry - invalid handle/queue. Looks as if it could be related to some manner of faulty hardware and/or connectivity.. > > Which logs do you need for debug it? Obtaining the full dmesg output via BMC ipmi or serial console would be very helpful to diagnose the issue. Also, qla2xxx + tcm_qla2xxx verbose debugging can be enabled via a module parameter, see this thread for more info: http://permalink.gmane.org/gmane.linux.scsi.target.devel/5136 And of course, please verify the latest ISP24xx firmware in place from http://ldriver.qlogic.com/firmware/ Thanks for reporting! --nab -- To unsubscribe from this list: send the line "unsubscribe target-devel" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
