Thanks for you tips, I would have a try as you said to disable the NOOP.
I had make a XFS file system on the LUN at the Initiator side .
Finally, I catch the sync_cache command was issued by XFS log
infrastructure actually. When I replace the XFS with EXT2 that dmesg
error don`t appear anymore
Hi Chris,
Thank you for you reply.
Just to be specific, can I get the scsi commands and try to do some process
on it in function like iscsi_data_xmit*(struct* iscsi_conn ***conn*) *before
the scsi commands are finally transmitted to the iscsi target?
Thanks
2016-05-20 18:18 GMT-05:00 Chris
On Mon, May 09, 2016 at 08:32:54AM -0700, whls.j...@gmail.com wrote:
> Hi,
>
> I am recently looking into the process of iSCSI initiator. I wonder where
> the source codes are that receive the scsi commands and encapsulate them
> into iscsi format. I have walked through the interaction between
On 05/20/2016 04:14 PM, Mike Christie wrote:
> On 05/20/2016 11:39 AM, The Lee-Man wrote:
>> Hi:
>>
>> It seems like your backend is getting busy and not replying in time when
>> it gets very busy. You can disable the NOOP, or you can lengthen its
>> interval, I believe.
>>
>> If there is a bug,
On 05/20/2016 11:39 AM, The Lee-Man wrote:
> Hi:
>
> It seems like your backend is getting busy and not replying in time when
> it gets very busy. You can disable the NOOP, or you can lengthen its
> interval, I believe.
>
> If there is a bug, it would be in the kernel target subsystem. Have you
Hi:
It seems like your backend is getting busy and not replying in time when it
gets very busy. You can disable the NOOP, or you can lengthen its interval,
I believe.
If there is a bug, it would be in the kernel target subsystem. Have you
tried the target-devel @ vger kernel mailing list?
On