On Tue, 2004-07-20 at 17:36, Hans de Bruin wrote:
> Hans de Bruin wrote:
> >
> > Message: 4
> > Subject: Re: [Iscsitarget-devel] an introduction
> > From: Ming Zhang <mingz@...>
> > Reply-To: mingz@...
> > To: Hans de Bruin <Bruinjm@...>
> > Cc: iscsitarget-devel@...
> > Date: Wed, 14 Jul 2004 17:44:23 -0400
> >
> > On Wed, 2004-07-14 at 16:19, Hans de Bruin wrote:
> >
> >> Hi there,
> >>
> >> A short introduction. I am a MS Windows admin specialized in SQL and
> >> SQL clustering. Since clustering hardware is kind of expensive i
> >> experiment using vmware. Since the last upgrade vmware is not suited
> >> for clustering any more. The shared disk have a high chance of getting
> >> corrupt during fail over.
> >>
> >> although iscsi et has not hit 1.0 yet i am eager to give it a try. So
> >> i have created some virtual machines: a dc (w2k3), two member servers
> >> (w2k3 with MS initiator 1.04a) and a linux slackware 10 iscsi machine.
> >> I also downloaded rfc3720.
> >
> > thx for u interest. if u like to provide us some feedback, we will
> > speedup the work to support ms clustering (cluster server or sql).
>
> That would be nice. There are a lot of windows would be iscsi clients out
> there.
after i cleanup the isns support work, i will work on this.
>
> >
> >>
> >> now i am stuck with several questions:
> >>
> >> - Is there a os free linux packet sniffer which understand iscsi packets?
> >
> > ethereal understand this. it has windows and linux version.
>
> nice, looks like sniffer pro on windows.
this one is powerful, it can explain the iscsi and isns packet to you.
:P
>
> >
> >>
> >> - Is there more logging than my /var/syslog?
> >
> > only here and /var/log/message.
>
> I noticed that iscsi_trgtd -fdddddddddddddd gives a nice result to.
yes, add d will add debug level.
>
> >
> >>
> >> - Why is copying a file to my iscsi disk much much faster than
> >> formatting the disk
> >> (http://www.xs4all.nl/~bruinjm/iscsi/2004-07-14-format.jpg)? The hba
> >> network may look like 10mb but is virtual and a file copy goes of the
> >> 10mb scale. There are a lot of reconnects going on usual ending in a
> >> failed -17. I then need to reboot the linux box (now there's a new
> >> experience).
> >
> > this is because the local caching. format disk will be a synchronous
> > procedure. increase the log level and try to post u log message to us.
>
> I will try to make a capture of a msi login and disk format, and put it
> somewhere
> on the net. How many d's do you need?
i think i need a log with more connect/reconnect, and -17 u mention. i
need not much on iscsi data traffic.
and with u long "d" parameter, u can send me this log as well.
>
> As far as i looked for the moment it looks like the target is the one that
> is closing the tcp/ip socket during a disk format from the initiator.
> Ethereal shows a 'scsi read request' followed by several 'scsi data in'
> frames ending with fin fin-ack ack. A few seconds later the initiator logs
> in again.
i remember once a time, i cannot partition and format on iscsi disk as
well, but later this problem gone in some version.
>
> gr
> Hans
>
>
>
> -------------------------------------------------------
> This SF.Net email is sponsored by BEA Weblogic Workshop
> FREE Java Enterprise J2EE developer tools!
> Get your free copy of BEA WebLogic Workshop 8.1 today.
> http://ads.osdn.com/?ad_id=4721&alloc_id=10040&op=click
> _______________________________________________
> Iscsitarget-devel mailing list
> Iscsitarget-devel@...
> https://lists.sourceforge.net/lists/listinfo/iscsitarget-devel
--
--------------------------------------------------
| Ming Zhang, PhD. Student
| Dept. of Electrical & Computer Engineering
| College of Engineering
| University of Rhode Island
| Kingston RI. 02881
| e-mail: mingz at ele.uri.edu
| Tel. (401) 874-2293
| Fax. (401) 782-6422
| http://www.ele.uri.edu/~mingz/
--------------------------------------------------
|