Slava Pestov wrote: >On Wed, Jan 21, 2015 at 3:04 AM, Stephen R. van den Berg <srb@xxxxxxx> wrote: >> Jan 21 11:21:40 ip144 kernel: ata11.00: exception Emask 0x0 SAct 0x7fffffff SErr 0x0 action 0x0 >> Jan 21 11:21:40 ip144 kernel: ata11.00: irq_stat 0x40000008 >> Jan 21 11:21:40 ip144 kernel: ata11.00: failed command: READ FPDMA QUEUED >> Jan 21 11:21:40 ip144 kernel: ata11.00: cmd 60/00:c0:10:d0:d9/04:00:01:00:00/40 tag 24 ncq 524288 in >> Jan 21 11:21:40 ip144 kernel: res 41/40:00:d0:d1:d9/00:00:01:00:00/00 Emask 0x409 (media error) <F> >> Jan 21 11:21:40 ip144 kernel: ata11.00: status: { DRDY ERR } >> Jan 21 11:21:40 ip144 kernel: ata11.00: error: { UNC } >> Jan 21 11:21:40 ip144 kernel: ata11.00: configured for UDMA/133 >I'm not sure this is related. Do you see it during normal operation >ever? It is possible that we're spinning in softirq context or >something, starving the device, but I'm not sure... In the second run now, with increased traffic, I do not see the above happening. Maybe I cut it short, the whole message is: Jan 21 12:09:14 ip144 kernel: ata11.00: exception Emask 0x0 SAct 0x3000 SErr 0x0 action 0x0 Jan 21 12:09:14 ip144 kernel: ata11.00: irq_stat 0x40000008 Jan 21 12:09:14 ip144 kernel: ata11.00: failed command: READ FPDMA QUEUED Jan 21 12:09:14 ip144 kernel: ata11.00: cmd 60/00:68:e8:2b:da/01:00:01:00:00/40 tag 13 ncq 131072 in Jan 21 12:09:14 ip144 kernel: res 41/40:00:b0:2c:da/00:00:01:00:00/00 Emask 0x409 (media error) <F> Jan 21 12:09:14 ip144 kernel: ata11.00: status: { DRDY ERR } Jan 21 12:09:14 ip144 kernel: ata11.00: error: { UNC } Jan 21 12:09:14 ip144 kernel: ata11.00: configured for UDMA/133 Jan 21 12:09:14 ip144 kernel: sd 10:0:0:0: [sdg] Jan 21 12:09:14 ip144 kernel: Result: hostbyte=0x00 driverbyte=0x08 Jan 21 12:09:14 ip144 kernel: sd 10:0:0:0: [sdg] Jan 21 12:09:14 ip144 kernel: Sense Key : 0x3 [current] [descriptor] Jan 21 12:09:14 ip144 kernel: Descriptor sense data with sense descriptors (in hex): Jan 21 12:09:14 ip144 kernel: 72 03 11 04 00 00 00 0c 00 0a 80 00 00 00 00 00 Jan 21 12:09:14 ip144 kernel: 01 da 2c b0 Jan 21 12:09:14 ip144 kernel: sd 10:0:0:0: [sdg] Jan 21 12:09:14 ip144 kernel: ASC=0x11 ASCQ=0x4 Jan 21 12:09:14 ip144 kernel: sd 10:0:0:0: [sdg] CDB: Jan 21 12:09:14 ip144 kernel: cdb[0]=0x88: 88 00 00 00 00 00 01 da 2b e8 00 00 01 00 00 00 Jan 21 12:09:14 ip144 kernel: blk_update_request: I/O error, dev sdg, sector 31075504 Jan 21 12:09:14 ip144 kernel: ata11: EH complete And it always happens upon a read error from HDD. -- Stephen. -- To unsubscribe from this list: send the line "unsubscribe linux-bcache" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html