On Wed, 2013-12-18 at 00:13 +0100, Henrik Goldman wrote: > Hello, > > Can someone explain what the following means? We see this from time to > time in our logs: > > [550122.418057] ABORT_TASK: Found referenced qla2xxx task_tag: 1217292 > [550124.066536] ABORT_TASK: Sending TMR_FUNCTION_COMPLETE for ref_tag: 1217292 > [550124.066551] ABORT_TASK: Sending TMR_TASK_DOES_NOT_EXIST for ref_tag: 1199736 > [550124.066558] ABORT_TASK: Sending TMR_TASK_DOES_NOT_EXIST for ref_tag: 1165460 > [550124.066563] ABORT_TASK: Sending TMR_TASK_DOES_NOT_EXIST for ref_tag: 1161896 > [550124.066566] ABORT_TASK: Sending TMR_TASK_DOES_NOT_EXIST for ref_tag: 1140644 > [550124.066569] ABORT_TASK: Sending TMR_TASK_DOES_NOT_EXIST for ref_tag: 1178616 > [550124.066573] ABORT_TASK: Sending TMR_TASK_DOES_NOT_EXIST for ref_tag: 1133032 > It means that a outstanding task (or number of tasks) have not generated a target response to the initiator port, causing an timeout to occur, and causing the initiator to issue an TMR ABORT_TASK to track down the outstanding I/O. This can be caused by a backend storage device taking a long time to process commands and generate a response. When using buffered FILEIO backends however, this typically would not be the case. Also, I've seen this type of timeout with older qla2xxx firmware. I'd recommend updating to latest at http://ldriver.qlogic.com/firmware/ on both initiator and target HBAs. The latest versions are: ql2400_fw.bin -- 7.01.00 MID ql2500_fw.bin -- 7.01.00 MIDQ --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
