Re: Problem connecting iscsi lun: DID_TRANSPORT_DISRUPTED...

2011-01-17 Thread Pierre-Yves Langlois
After issuing the command  echo 1 
/sys/module/libiscsi/parameters/debug_libiscsi_eh,I didn't see any changes
in
/var/log/messages, /var/log/kern.log, /var/log/daemon.log, /var/log/debug.
Do I look at the right place?


On Sat, Jan 15, 2011 at 1:54 AM, Mike Christie micha...@cs.wisc.edu wrote:

 On 01/14/2011 02:40 PM, PYL wrote:

 I'm not able to connect to a iscsi lun. I use proxmox ve 1.7 but I
 have installed the kernel 2.6.36.2 to fix a bug with my network card
 drivers...

 What am I missing?

 Here is the output from /var/log/messages:

 [CODE]
 Jan 14 14:57:13 fl-vm01 kernel: vmbr1: received packet on bond0.222
 with own address as source address
 Jan 14 14:57:13 fl-vm01 kernel: scsi3 : iSCSI Initiator over TCP/IP
 Jan 14 14:57:13 fl-vm01 kernel: scsi 3:0:0:0: Direct-Access
 NETAPP   LUN  7330 PQ: 0 ANSI: 4
 Jan 14 14:57:13 fl-vm01 kernel: sd 3:0:0:0: Attached scsi generic sg3
 type 0
 Jan 14 14:57:13 fl-vm01 kernel: sd 3:0:0:0: [sdb] 1258450944 512-byte
 logical blocks: (644 GB/600 GiB)
 Jan 14 14:57:13 fl-vm01 kernel: sd 3:0:0:0: [sdb] Write Protect is off
 Jan 14 14:57:13 fl-vm01 kernel: sd 3:0:0:0: [sdb] Write cache:
 disabled, read cache: enabled, doesn't support DPO or FUA
 Jan 14 14:57:23 fl-vm01 kernel: connection1:0: detected conn error
 (1011)
 Jan 14 14:57:37 fl-vm01 kernel: connection1:0: detected conn error
 (1011)
 Jan 14 14:57:52 fl-vm01 kernel: connection1:0: detected conn error
 (1011)
 Jan 14 14:58:07 fl-vm01 kernel: connection1:0: detected conn error
 (1011)


 We seem to be getting them around every 15 seconds, so I think a scsi
 command is timing out which is starting the scsi eh and we end up dropping
 the session and relogging in because TMFs do not work.

 You could do

 echo 1  /sys/module/libiscsi/parameters/debug_libiscsi_eh

 then rerun your test to confirm this.

 I do not know why the command is timing out though. Is there anything in
 the target logs? If you create a smaller LU (just a couple gigs) does it
 work then?


-- 
You received this message because you are subscribed to the Google Groups 
open-iscsi group.
To post to this group, send email to open-iscsi@googlegroups.com.
To unsubscribe from this group, send email to 
open-iscsi+unsubscr...@googlegroups.com.
For more options, visit this group at 
http://groups.google.com/group/open-iscsi?hl=en.



Re: Problem connecting iscsi lun: DID_TRANSPORT_DISRUPTED...

2011-01-17 Thread Mike Christie

On 01/17/2011 02:26 PM, Pierre-Yves Langlois wrote:

After issuing the command  echo 1
/sys/module/libiscsi/parameters/debug_libiscsi_eh,I didn't see any changes
in
/var/log/messages, /var/log/kern.log, /var/log/daemon.log, /var/log/debug.
Do I look at the right place?



Did you do the echo, then relogin to the target?

echo 1  /sys/module/libiscsi/parameters/debug_libiscsi_eh
iscsiadm -m node -u
iscsiadm -m node -l

--
You received this message because you are subscribed to the Google Groups 
open-iscsi group.
To post to this group, send email to open-iscsi@googlegroups.com.
To unsubscribe from this group, send email to 
open-iscsi+unsubscr...@googlegroups.com.
For more options, visit this group at 
http://groups.google.com/group/open-iscsi?hl=en.



DCBX support

2011-01-17 Thread Rustad, Mark D
Hi all,

I am still looking into DCBX support for iSCSI and potentially other networking 
applications. Previously I was only considering having the application be a 
client of the lldpad daemon, but it looks like the application priority 
information should be available from rtnetlink. Since rtnetlink is already 
often used by networking applications, it seems like a better fit to use that 
interface.

Right now, the only missing element seems to be an rtnetlink event that could 
be used to trigger the app to change priorities based on a change in the 
configuration. How important do you think it is to support dynamic changes to 
application priority? Is it essential for a first implementation?

I think that this approach results in better package and code dependencies. 
Many networking apps already rely on the rtnetlink interface, so this would be 
a natural extension of that.

Thanks to Eric Multanen for suggesting this approach.

-- 
Mark Rustad, mark.d.rus...@intel.com

-- 
You received this message because you are subscribed to the Google Groups 
open-iscsi group.
To post to this group, send email to open-iscsi@googlegroups.com.
To unsubscribe from this group, send email to 
open-iscsi+unsubscr...@googlegroups.com.
For more options, visit this group at 
http://groups.google.com/group/open-iscsi?hl=en.