On Thu, Apr 01, 2021 at 07:27:10AM +0200, Bruno Flueckiger wrote:
> On 31.03., David Alten wrote:
> > Hello,
> >
> > I???m having issues getting iscsid to work with my Synology NAS.
> >
> > The first issue was that the NAS was returning an error code. Turns out
> > it didn???t like not?? missing the default SessionType=Normal.
> >
> > The second issue was that the login sequence never comleted. It seems we
> > keep hardcoding MaxConnections and MaxRecvDataSegmentLength in
> > connection.c:conn_gen_kvp() instead of negotiating to the lowest value.
> > So I just bypassed that function entirely and hardcoded all the default
> > settings in initiator.c.
> >
> > This makes the login step complete, but then nothing else happens after
> > that. I would expect a new scsi device to be attached so I could use it.
> >
> > I did notice that MaxConnection shows as 0, but the NAS does show the
> > connection as established.
> >
> > b1# iscsictl show?? ??
> > Initiator: ISID base 80d3cf6f qalifier 6e7d
> >
> > Session 'disk2':
> > ?? ?? SessionType: normal MaxConnections: 0
> > ?? ?? TargetName: iqn.2000-01.com.synology:Target.02
> > ?? ?? TargetAddr: 192.168.0.4:iscsi
> > ?? ?? InitiatorName: iqn.1995-11.org.openbsd.iscsid:b1
> > ?? ?? InitiatorAddr: 192.168.0.9
> > command successful
> >
> > b1$ cat /etc/iscsi.conf?? ?? ?? ?? ?? ?? ?? ?? ?? ?? ?? ?? ?? ?? ?? ?? ??
> > ?? ?? ?? ?? ?? ?? ?? ??
> > ?? ?? ?? ?? ?? ?? ?? ?? ?? ?? ?? ?? ?? ?? ?? ?? ?? ?? ?? ?? ????
> > target disk2 {
> > ?? ?? ?? ?? initiatoraddr 192.168.0.9
> > ?? ?? ?? ?? targetaddr 192.168.0.4
> > ?? ?? ?? ?? targetname "iqn.2000-01.com.synology:Target.02"
> > }
> > Any suggestions on where to go from here?
> >
> > Thanks,
> >
> > -David
> >
>
> Hi David,
>
> I had the same problem some years ago when I tried to attach an iSCSI
> LUN from my Synology to my APU box. I wrote the patch below which fixed
> the problem for me. Maybe you can give it a try?
>
Can someone send me a full pcap file of the negotiation with the synology
nas device? It seems there is something not quite right in the negotiation
code but the proposed diffs don't look correct either.
--
:wq Claudio
