Re: [Iscsitarget-devel] a question about max_cmd_sn

[Ross S. W. W. <RW...@me...>]

spren [mailto:sp...@16...] wrote:
> 
> Hi Ross,
> 
> I have a question about the max_queued_cmnds and the 
> max_cmd_sn in iet. it seems that the max_cmd_sn in iscsi response
> PDU is always set to (sess->exp_cmd_sn + 
> sess->max_queued_cmnds) by kernel/iscsi.c/cmnd_set_sn(), and 
> max_queued_cmnds
> is a relative fixed positive value, may this lead to an 
> endless write command and DATA-OUT received from initiators 
> and consuming too much memory in iet ? if writing process is slow.

Good observation, and yes, you are correct, the current behavior
gives a limitless command buffer, which can put memory pressure
on the kernel, but we usually see the pressure more on the disk
subsystem where frequent SCSI timeouts start occurring and then
frequent cmnd_aborts start occurring.

I have a fix for this, but it wasn't committed to 1.4 it will
most likely appear in 1.5 along with several other big fixes,
(proper tagged command queuing, PGRs, to name a few).

If you are experiencing these issues let me know and I can
provide a patch to test out to see if it fixes things for
you.

-Ross

______________________________________________________________________
This e-mail, and any attachments thereto, is intended only for use by
the addressee(s) named herein and may contain legally privileged
and/or confidential information. If you are not the intended recipient
of this e-mail, you are hereby notified that any dissemination,
distribution or copying of this e-mail, and any attachments thereto,
is strictly prohibited. If you have received this e-mail in error,
please immediately notify the sender and permanently delete the
original and any copy or printout thereof.