continuosly change TCP port of starport

[tufillaro]

Good Morning
I have this problem
Configuration
two server windows 2003 RC 2 (but it happens also in all configuratione XP svc pck2 windows svc pack 1 etc.)
the network connection is on 1 Gbyte Realtek board cross cable direct (UTP 6)

One hosts starwind the other one starport or MSCSI (but not in the same config)

I set the starwind machine for using native SATA controller or IDE or SCSI or RAM
No problem in the starport machine it works, the performance are accettable etc.

I set the starwind machine for using extra SATA controller (PCI based) and it works very well in the server machine .

In the starport machine the CPU repeatly go to 100% the machine hangs and after a period (2 sec) go to %0 and the machine not in hang.

Monitoring by netstat command I see that the starport changes continuosly .
Crossing control with the log file on the starwind machine the log file tell that
' an old session on the port ... ' it is now reconnected on the other port

the result is that in the starport machine i see with very difficult a mountvol but it is false and every action vs falls.
All the same actions NOT fail in the other configs described below.

Some suggestions ?

Thanks in advance.

[Val (staff)]

Hi,

Could you send me the StarWind's log file?
This is a known issue when "Session reinstatement starting" message is logged. Some times the old session can't be closed in the right way and any new sessions for the target are rejected.
We are working on the problem right now, it will be fixed in the next minor release of StarWind.

As a work around you could try to mark the StarWind target as 'Clustered'. But be careful not to mount the target from 2+ machines simultaneously, because this could break the logical structure of the volume.

[tufillaro]

Ok

PS:
I don't found clustered
I have a SATA controller (two disks) NOT RAID setted
How could I realize your suggestion ?

*** begin attach
6/30 7:43:21 23c Srv: StarWind iSCSI Target v3.1.1 (Build 20060531, Win32)

6/30 7:43:21 23c Srv: Built May 31 2006 15:23:24
...

[Val (staff)]

Ok

PS:
I don't found clustered
I have a SATA controller (two disks) NOT RAID setted
How could I realize your suggestion ?

*** begin attach
6/30 7:43:21 23c Srv: StarWind iSCSI Target v3.1.1 (Build 20060531, Win32)

6/30 7:43:21 23c Srv: Built May 31 2006 15:23:24
...

Please don't copy&paste huge logs into your messages. Size of a message can't be more than 64K, so your logs are truncated.

Please send me the log file zipped to my email: valery@rocketdivision.com

For SPTI devices you need use 'sessions' parameter instead of 'clustering' to allow more active sessions to the device.

[Val (staff)]

Thank you for the log.

1) Have you tried the MS iSCSI initiator with the same target?
Are there any differences?
The log file includes no error messages, it gives us no idea why the StarPort initiator keeps trying to reestablish the session while the previous one seems to be OK...
Are there any Ethereal log files for the session?

2) Have you tried the ImageFile plug-in to create virtual targets instead of direct mapping of hard drives?
(This is a recommended way to provide storage to the clients with StarWind iSCSI target)