iscsi initiator troubles

[John Jaser <john%jaser.net@localhost>]

Hi,



I've been struggling for a few days getting iscsi_initiator talking to a
netbsd 5.1 iscsi-target. The iscsi-target has been performing well for about a
solid week as a back-end for a debian 6.0 xen server using open-iscsi. Getting
a netbsd 5.1 xen server to do the same has been the challenge. All systems are
amd64 with 8G RAM, using amd64 kernels.



The symptom is not apparent until file creation, but happens rather quickly.
The file copy or dd empty-file creation will just hang, or sometimes barely
get started. The system becomes unresponsive. The 'top' process list in
another terminal continues to run, and shows almost no load (too low) even on
the iscsi-initiator process and the dd or cp process. New terminal logins
work, but any command other than an 'ls' hangs also, and the situation
deteriorates until a dirty restart is required. I also tried using
iscsi_initiator on the iscsi-target system with similar results. Lastly, I
tried enabling iscsi-target on the iscsi_initiator host itself, and got the
same result. That session is below. Of interest may be the two errors in the
syslog followed by the successful login. I tried the 5.1-XEN kernel(with
puffs), the generic 5.1 kernel, and a current kernel from last week; all gave
same result. Seems almost like some kind of deadlock, since the load is so
low. Any tips would be greatly appreciated.


cheers,


JJ


------------------------------



syslog messages:


Sep 25 17:10:09 xen1 iscsi-target: pid
96:/home/builds/ab/netbsd-5-1-RELEASE/src/dist/iscsi/src/target.c:932:
***ERROR*** cmd.tsih 0 not found
Sep 25 17:10:09 xen1 iscsi-target: pid
96:/home/builds/ab/netbsd-5-1-RELEASE/src/dist/iscsi/src/target.c:911:
***ERROR*** Abnormal SessionType cmd.tsih 0 not found
Sep 25 17:10:09 xen1 iscsi-target: > iSCSI Discovery login successful from
iqn.1994-04.org.NetBSD.iscsi-initiator:agc on 127.0.0.1 disk 0, ISID 0, TSIH
11
Sep 25 17:10:09 xen1 iscsi-target: < iSCSI Discovery logout successful from
iqn.1994-04.org.NetBSD.iscsi-initiator:agc on 127.0.0.1 disk 0, ISID 0, TSIH
11
Sep 25 17:10:09 xen1 iscsi-target: > iSCSI Normal login successful from
iqn.1994-04.org.NetBSD.iscsi-initiator:agc on 127.0.0.1 disk 0, ISID 0, TSIH
12


--------------------------------



xen1# cat /etc/iscsi/targets


# extent file or device start length
extent0 /iscsi_t/iscsi-target0 0 1000MB


# target flags storage netmask
target0 rw extent0 0.0.0.0/0


xen1# /etc/rc.d/iscsi_target forcestart
Starting iscsi_target.
Reading configuration from `/etc/iscsi/targets'
target0:rw:0.0.0.0/0
extent0:/iscsi_t/iscsi-target0:0:1048576000
DISK: 1 logical unit (2048000 blocks, 512 bytes/block), type iscsi fs
DISK: LUN 0: 1000 MB disk storage for "target0"
TARGET: TargetName is iqn.1994-04.org.netbsd.iscsi-target
xen1# iscsi-initiator -D -h 127.0.0.1
Targets available from host xen1.lightpacket.com:
iqn.1994-04.org.netbsd.iscsi-target:target0 at 127.0.0.1:3260,1
xen1# iscsi-initiator -h 127.0.0.1 /iscsi &
xen1# bg
[1] iscsi-initiator -h 127.0.0.1 /iscsi
xen1# pgrep iscsi
485
475
xen1# ps 485
PID TTY STAT TIME COMMAND
485 ttyp0 Il 0:00.00 iscsi-initiator -h 127.0.0.1 /iscsi
xen1# ps 475
PID TTY STAT TIME COMMAND
475 ? Isl 0:00.00 /usr/sbin/iscsi-target
xen1# ls -l /iscsi/target0
total 448
lrw-r--r-- 1 root wheel 9 Sep 25 15:08 hostname -> 127.0.0.1
lrw-r--r-- 1 root wheel 9 Sep 25 15:08 ip -> 127.0.0.1
lrw-r--r-- 1 root wheel 16 Sep 25 15:08 product -> NetBSD iSCSI
- rw-r--r-- 1 root wheel 1048576000 Sep 25 15:08 storage
lrw-r--r-- 1 root wheel 43 Sep 25 15:08 targetname ->
iqn.1994-04.org.netbsd.iscsi-target:target0
lrw-r--r-- 1 root wheel 8 Sep 25 15:08 vendor -> NetBSD
lrw-r--r-- 1 root wheel 4 Sep 25 15:08 version -> 0
xen1# vnconfig vnd0 /iscsi/target0/storage
xen1# disklabel vnd0
# / dev/rvnd0d:
type: vnd
disk: vnd
label: fictitious
flags:
bytes/sector: 512
sectors/track: 32
tracks/cylinder: 64
sectors/cylinder: 2048
cylinders: 1000
total sectors: 2048000
rpm: 3600
interleave: 1
trackskew: 0
cylinderskew: 0
headswitch: 0 # microseconds
track-to-track seek: 0 # microseconds
drivedata: 0


4 partitions:
# size offset fstype [fsize bsize cpg/sgs]
a: 2048000 0 4.2BSD 0 0 0 # (Cyl. 0 - 999)
d: 2048000 0 unused 0 0 # (Cyl. 0 - 999)
xen1# newfs -O 1 /dev/rvnd0a
/ dev/rvnd0a: 1000.0MB (2048000 sectors) block size 16384, fragment size 2048
using 6 cylinder groups of 166.67MB, 10667 blks, 20992 inodes.
super-block backups (for fsck_ffs -b #) at:
32, 341376, 682720, 1024064, 1365408, 1706752,
xen1# fsck /dev/vnd0a
** / dev/rvnd0a
** File system is clean; not checking
xen1# mount /dev/vnd0a /testjj
xen1# cd /testjj
xen1# df -h
Filesystem Size Used Avail %Cap Mounted on
/ dev/wd0a 917G 357G 514G 40% /
kernfs 1.0K 1.0K 0B 100% /kern
ptyfs 1.0K 1.0K 0B 100% /dev/pts
procfs 4.0K 4.0K 0B 100% /proc
/ dev/puffs 0B 0B 0B 100% /iscsi
/ dev/vnd0a 984M 2.0K 935M 0% /testjj
xen1# mdkir test
mdkir: not found
xen1# mkdir test
xen1# cd test
xen1# ls -la
total 4
drwxr-xr-x 2 root wheel 512 Sep 25 16:44 .
drwxr-xr-x 3 root wheel 512 Sep 25 16:44 ..


xen1# cp /home/john/iso/i386cd-5.1.iso .
xen1# ls
i386cd-5.1.iso
xen1# diff i386cd-5.1.iso /home/john/iso/i386cd-5.1.iso
xen1# df -h
Filesystem Size Used Avail %Cap Mounted on
/ dev/wd0a 917G 357G 514G 40% /
kernfs 1.0K 1.0K 0B 100% /kern
ptyfs 1.0K 1.0K 0B 100% /dev/pts
procfs 4.0K 4.0K 0B 100% /proc
/ dev/puffs 0B 0B 0B 100% /iscsi
/ dev/vnd0a 984M 238M 697M 25% /testjj
xen1# ls
i386cd-5.1.iso
xen1# pwd
/ testjj/test
xen1# cp *.iso 2.iso


(system hangs here)