On 1/4/14 8:08 AM, [email protected] wrote:
Hello,
when trying to connect open-iscsi in initiator mode to solaris target I
have encountered situation when accessing of iscsi device causes hang on
kernel call.
iscsi successfully logs on to targets, devices /dev/sdc and /dev/sdd (one
for IPv4 and another for IPv6) are created, by trying to partition it
causes partition program to hang:
[ 3960.444075] INFO: task parted:4819 blocked for more than 120 seconds.
[ 3960.444079] "echo 0 > /proc/sys/kernel/hung_task_timeout_secs" disables
this message.
[ 3960.444081] parted D ffff88007a8d4ba0 0 4819 1
0x00000004
[ 3960.444086] ffff88007a8d4880 0000000000000086 0000000000014300
ffff88005a713fd8
[ 3960.444090] 0000000000014300 ffff88005a713fd8 ffff88007dc94b38
ffff88007de8af48
[ 3960.444093] 0000000000000002 ffffffff8110afb0 ffff88005a713d60
0000000000c7ffff
[ 3960.444097] Call Trace:
[ 3960.444105] [<ffffffff8110afb0>] ? wait_on_page_read+0x60/0x60
[ 3960.444110] [<ffffffff81477e64>] ? io_schedule+0x94/0x120
[ 3960.444113] [<ffffffff8110afb5>] ? sleep_on_page+0x5/0x10
[ 3960.444116] [<ffffffff81475ea4>] ? __wait_on_bit+0x54/0x80
[ 3960.444120] [<ffffffff8110adcf>] ? wait_on_page_bit+0x7f/0x90
[ 3960.444124] [<ffffffff81078b90>] ? wake_atomic_t_function+0x30/0x30
[ 3960.444128] [<ffffffff81117318>] ? pagevec_lookup_tag+0x18/0x20
[ 3960.444131] [<ffffffff8110aeb8>] ? filemap_fdatawait_range+0xd8/0x150
[ 3960.444134] [<ffffffff81114f7f>] ? generic_writepages+0x3f/0x50
[ 3960.444138] [<ffffffff8110c421>] ? filemap_write_and_wait+0x31/0x50
[ 3960.444142] [<ffffffff8123aaef>] ? blkdev_ioctl+0x42f/0x7c0
[ 3960.444145] [<ffffffff8119d9a8>] ? block_ioctl+0x38/0x40
[ 3960.444148] [<ffffffff8117ae54>] ? do_vfs_ioctl+0x2d4/0x4b0
[ 3960.444152] [<ffffffff81169cd4>] ? vfs_write+0x154/0x1e0
[ 3960.444155] [<ffffffff8119dab9>] ? block_llseek+0x49/0x60
[ 3960.444158] [<ffffffff8117b0b0>] ? SyS_ioctl+0x80/0xa0
[ 3960.444161] [<ffffffff8116a643>] ? SyS_write+0x43/0xa0
[ 3960.444165] [<ffffffff81481469>] ? system_call_fastpath+0x16/0x1b
Writing to device with dd have same effect. Hanged process cannot be killed
and system cannot be shutted down gracefully.
The only rescue is to reboot with Ctrl-Alt-SysRq-b
When the hang occurs, excessive network traffic starts. tcpdump shows
numerous packets containing zeroes flooding on iscsi connection to target
host.
You would have to send the complete /var/log/messages, and also send the
tcpdump trace.
What do you mean by broken target? What is broken about it? Is it slowly
responding to IO requests by any chance?
--
You received this message because you are subscribed to the Google Groups
"open-iscsi" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to [email protected].
To post to this group, send email to [email protected].
Visit this group at http://groups.google.com/group/open-iscsi.
For more options, visit https://groups.google.com/groups/opt_out.
