[dm-devel] scsi errors

[Guy Coates]

I've got two test systems here, (IA64 and IA32)  and am seeing the same
behaviour, but only on the IA32 system.

The setup is as follows: both machines are running 2.6.10+udm2, with dual
QLA2312 cards (the IA64 has rev 15 and the IA32 system rev 02 cards).
Both machines are connected via a brocade fabric to a dual-controller
Compaq HSV110. There are 10 luns presented from the HSV110 to each
machine.

I'm using the following multipath conf for the devices:

path_selector           "round-robin 0"
path_grouping_policy "group_by_serial"
getuid_callout "/sbin/scsi_id -g -u -s /block/%n"
path_checker "tur"
features "1 queue_if_no_path"


The active device maps are a single logical volume:

hsv110lun1-mysqltest1: 0 209584128 linear

consisting of a single multipath physical volume:

3600508b40001417e0001c00001b60000: 0 209715200 multipath 1 0 0 2 2 E 0 2 0 65:240 A 0 8:176 A 0 A 0 2 0 8:16 A 0 65:80 A 1

The filesystem is XFS.

To trigger the problem I simply mount the filesystem and wait. There is no
IO  apart from the filesystem journaling activity.

Jan  7 23:49:55 web-4-09 kernel: XFS mounting filesystem dm-11

<No errors reported>

Jan  8 01:38:51 web-4-09 kernel: scsi(5): Found in Pending queue pid 468267, State = 6., fcport state=4 jiffies=703298
Jan  8 01:38:51 web-4-09 kernel: scsi(5:0:6) qla2x00_done: did_error = 2, comp-scsi= 0x0-0x0.
Jan  8 01:39:08 web-4-09 kernel: scsi(5): Found in Pending queue pid 468268, State = 6., fcport state=4 jiffies=707500
Jan  8 01:39:08 web-4-09 kernel: scsi(5:0:6) qla2x00_done: did_error = 2, comp-scsi= 0x0-0x0.
Jan  8 01:39:25 web-4-09 kernel: scsi(5): Found in Pending queue pid 468269, State = 6., fcport state=4 jiffies=70b768
Jan  8 01:39:25 web-4-09 kernel: scsi(5:0:6) qla2x00_done: did_error = 2, comp-scsi= 0x0-0x0.
Jan  8 01:39:42 web-4-09 kernel: scsi(5): Found in Pending queue pid 468270, State = 6., fcport state=4 jiffies=70f9d0
Jan  8 01:39:42 web-4-09 kernel: scsi(5:0:6) qla2x00_done: did_error = 2, comp-scsi= 0x0-0x0.
Jan  8 01:39:59 web-4-09 kernel: scsi(5): Found in Pending queue pid 468271, State = 6., fcport state=4 jiffies=713c38
Jan  8 01:39:59 web-4-09 kernel: scsi(5:0:6) qla2x00_done: did_error = 2, comp-scsi= 0x0-0x0.
Jan  8 01:40:16 web-4-09 kernel: scsi(5): Found in Pending queue pid 468272, State = 6., fcport state=4 jiffies=717ea0
Jan  8 01:40:16 web-4-09 kernel: scsi(5:0:6) qla2x00_done: did_error = 2, comp-scsi= 0x0-0x0.

<snip>

Jan  8 06:50:50 web-4-09 kernel: scsi(4:0:1) qla2x00_done: did_error = 2, comp-scsi= 0x0-0x0.
Jan  8 06:50:51 web-4-09 kernel: scsi(4): Found in Pending queue pid 469587, State = 6., fcport state=4 jiffies=18de3ee
Jan  8 06:50:51 web-4-09 kernel: scsi(4:0:1) qla2x00_done: did_error = 2, comp-scsi= 0x0-0x0.
Jan  8 06:50:53 web-4-09 kernel: scsi(4): Found in Pending queue pid 469589, State = 6., fcport state=4 jiffies=18debbf
Jan  8 06:50:53 web-4-09 kernel: scsi(4:0:1) qla2x00_done: did_error = 2, comp-scsi= 0x0-0x0.
Jan  8 06:50:55 web-4-09 kernel: scsi(4): Found in Pending queue pid 469590, State = 6., fcport state=4 jiffies=18df38f
Jan  8 06:50:55 web-4-09 kernel: scsi(4:0:1) qla2x00_done: did_error = 2, comp-scsi= 0x0-0x0.
Jan  8 06:50:57 web-4-09 kernel: scsi(4): Found in Pending queue pid 469591, State = 6., fcport state=4 jiffies=18dfb5f
Jan  8 06:50:57 web-4-09 kernel: scsi(4:0:1) qla2x00_done: did_error = 2, comp-scsi= 0x0-0x0.
Jan  8 06:50:57 web-4-09 kernel: SCSI error : <4 0 0 1> return code = 0x20000
Jan  8 06:50:57 web-4-09 kernel: scsi(4:0:2) UNDERRUN status detected 0x15-0x800.

<snip>

Jan  8 10:36:17 web-4-09 kernel: scsi(5): Found in Pending queue pid 531196, State = 6., fcport state=4 jiffies=25c5130
Jan  8 10:36:17 web-4-09 kernel: scsi(5:0:1) qla2x00_done: did_error = 2, comp-scsi= 0x0-0x0.
Jan  8 10:36:17 web-4-09 kernel: SCSI error : <5 0 0 1> return code = 0x20000
Jan  8 10:36:17 web-4-09 kernel: end_request: I/O error, dev sdv, sector 65920
Jan  8 10:36:17 web-4-09 kernel: scsi(5:0:6) UNDERRUN status detected 0x15-0x800.
Jan  8 10:36:27 web-4-09 kernel: scsi(5): Found in Pending queue pid 531198, State = 6., fcport state=4 jiffies=25c75af
Jan  8 10:36:27 web-4-09 kernel: scsi(5:0:6) qla2x00_done: did_error = 2, comp-scsi= 0x0-0x0.



Cheers,

Guy Coates

-- 
Dr. Guy Coates,  Informatics System Group
The Wellcome Trust Sanger Institute, Hinxton, Cambridge, CB10 1SA, UK
Tel: +44 (0)1223 834244 ex 7199