Menu
▾
▴
#24 Failing command cdb 0x28
Hello, I was just wondering if anyone could possibly
shed some light on a problem I've encountered. On a
fedora core 3 box, running kernel 2.6.10 with the 4.0.2
linux iscsi kernel patch and user utilities, I start up
iscsi and an iscsi-ls shows:
*******************************************************************************
SFNet iSCSI Driver Version ...4:0.2(15-Mar-2005)
*******************************************************************************
TARGET NAME :
iqn.2002-10.com.infortrend:raid.sn6780480.00
TARGET ALIAS : <NULL>
HOST NO : 16
TARGET ADDRESS : 10.0.0.60:3260,1
SESSION STATUS : CONNECTION ESTABLISHED AT Wed
Jul 13 20:58:18 BST 2005
SESSION ID : ISID 00023d000001
TSIH 200
-------------------------------------------------------------------------------
So that looks reasonable. However, in
/var/log/messages, I see lots of e.g.:
iscsi-sfnet:host16: Failing command cdb 0x28 task 57
with return code = 0x20000
If I try and do anything with the device, it's
behaviour is very flaky. Does anyone know what this
fail message means in practical terms, and how I should
go about resolving it? Any help mugh appreciated.
Thanks!
Logged In: NO
Could you post your whole log. It looks like commands are
failing becuase they timed out or the session was reset.
Make sure your network connection is good.
Logged In: YES
user_id=1120997
Thanks for your reply. Here are the logs resulting from
/etc/init.d/iscsi start:
Jul 14 09:16:54 kernel: Synchronizing SCSI cache for disk sda:
Jul 14 09:16:54 kernel: iscsi-sfnet:host16: Session dropped
Jul 14 09:16:58 iscsid[12636]: version 4:0.2 variant
(15-Mar-2005)
Jul 14 09:16:58 iscsid[12640]: Connected to Discovery
Address 10.0.0.60
Jul 14 09:16:58 kernel: iscsi-sfnet:host17: Session established
Jul 14 09:16:58 kernel: scsi17 : SFNet iSCSI driver
Jul 14 09:16:58 kernel: Vendor: IFT Model:
A12E-G2121-2 Rev: 342D
Jul 14 09:16:58 kernel: Type: Direct-Access
ANSI SCSI revision: 03
Jul 14 09:16:58 kernel: SCSI device sda: 1048576 512-byte
hdwr sectors (537 MB)
Jul 14 09:16:58 kernel: iscsi-sfnet:host17: recv_sg_data -
invalid sglist for offset 0 len 20, remaining data 12,
sglist size 1, dropping session
Jul 14 09:16:58 kernel: SCSI device sda: drive cache: write back
Jul 14 09:16:58 kernel: iscsi-sfnet:host17: iscsi_recv_pdu()
cannot read 16384 bytes of PDU data, only 4096 bytes of
buffer available
Jul 14 09:16:58 kernel: iscsi-sfnet:host17: Session has
ended quickly 1 times, login delay 2 seconds
Jul 14 09:16:58 kernel: iscsi-sfnet:host17: Session dropped
Jul 14 09:16:58 kernel: iscsi-sfnet:host17: Failing command
cdb 0x00 task 8 with return code = 0x20000
Jul 14 09:16:58 kernel: iscsi-sfnet:host17: Waiting 2
seconds before next login attempt
Jul 14 09:17:00 kernel: iscsi-sfnet:host17: Session established
Jul 14 09:17:00 kernel: SCSI device sda: 1048576 512-byte
hdwr sectors (537 MB)
Jul 14 09:17:00 kernel: iscsi-sfnet:host17: recv_sg_data -
invalid sglist for offset 0 len 20, remaining data 12,
sglist size 1, dropping session
Jul 14 09:17:00 kernel: SCSI device sda: drive cache: write back
Jul 14 09:17:00 kernel: sda:<3>iscsi-sfnet:host17:
iscsi_recv_pdu() cannot read 16384 bytes of PDU data, only
4096 bytes of buffer available
Jul 14 09:17:00 kernel: iscsi-sfnet:host17: Session has
ended quickly 2 times, login delay 2 seconds
Jul 14 09:17:00 kernel: iscsi-sfnet:host17: Session dropped
Jul 14 09:17:00 kernel: iscsi-sfnet:host17: Failing command
cdb 0x28 task 14 with return code = 0x20000
Jul 14 09:17:00 kernel: iscsi-sfnet:host17: Waiting 2
seconds before next login attempt
Jul 14 09:17:02 kernel: iscsi-sfnet:host17: Session established
Jul 14 09:17:02 kernel: sda1
Jul 14 09:17:02 kernel: Attached scsi disk sda at scsi17,
channel 0, id 0, lun 0
Jul 14 09:17:02 scsi.agent[12668]: disk at
/devices/platform/host17/target17:0:0/17:0:0:0
Logged In: NO
eek. What target are you using?
Logged In: YES
user_id=1120997
Thanks for your reply, the target is an Infortrend EonStor
A12E-G2121 iscsi 4TB RAID storage device, .
Logged In: YES
user_id=1120997
Does 'eek' roughly translate as I lose, then? (Shuddering at
the thought of 4TB of NTFS and samba).
Logged In: NO
No, it just means it is not a trivial bug so I will not get
to this week. I am about to get on plane to Canada where I
will be for the next week without a computer. I will get to
this bug when I get back though.
Logged In: YES
user_id=1120997
Just for some extra info, the problems I reported here seem
to have gone away with linux-iscsi-5.0.0.0.3rc6-363 built
against a 2.6.11.1 kernel, so I'll just go with the
experimental driver for now...