NETAPP LUN - strange error

[Maaz]

iscsi-initiator OS: SLES 10 SP 1 x86_64

we assigned 1024 MB LUN to the sles10. then we configure the sles10 as iscsi-initiator, not problem every thing went smooth.

but
# lsscsi
[11:0:0:0] no dev NETAPP LUN 0.2 -

in the 2nd column of output you can see "no dev" instead of disk, and the last column is "-" instead of /dev/sdX.

# tail -f /var/log/messages

Sep 11 13:28:57 backup002 kernel: Loading iSCSI transport class v2.0-754.
Sep 11 13:28:57 backup002 kernel: iscsi: registered transport (tcp)
Sep 11 13:28:57 backup002 iscsid: iSCSI logger with pid=12020 started!
Sep 11 13:28:57 backup002 kernel: scsi11 : iSCSI Initiator over TCP/IP
Sep 11 13:28:57 backup002 kernel: scsi: unknown device type 31
Sep 11 13:28:57 backup002 kernel: Vendor: NETAPP Model: LUN Rev: 0.2
Sep 11 13:28:57 backup002 kernel: Type: Unknown ANSI SCSI revision: 04
Sep 11 13:28:57 backup002 kernel: 11:0:0:0: Attached scsi generic sg0 type 31

*****following line is strange*****
Sep 11 13:28:57 backup002 kernel: scsi: host 11 channel 0 id 0 lun2340 has a LUN larger than allowed by the host adapter


Sep 11 13:28:58 backup002 iscsid: transport class version 2.0-754. iscsid version 2.0-754
Sep 11 13:28:58 backup002 iscsid: an InitiatorAlias= is required, but was not found in /etc/iscsi/initiatorname.iscsi
Sep 11 13:28:58 backup002 iscsid: iSCSI sync pid=12022 started
Sep 11 13:28:58 backup002 iscsid: iSCSI daemon with pid=12021 started!
Sep 11 13:28:58 backup002 iscsid: connection0:0 is operational now

What I did is that I disconnect the san from iscsi-initiator, delete the 1 GB LUN from NETAPP, re-create/assign a LUN of 7 GB to sles 10... and problem gone
now

# lssci
[12:0:0:20] disk NETAPP LUN 0.2 /dev/sda
and no strange messages in /var/log/messages

please comment/recommend/suggest/explain

what does following message means
host 11 channel 0 id 0 lun2340 has a LUN larger than allowed by the host adapter

Regards
Maaz

[Uwe Zessin]

> Sep 11 13:28:57 backup002 kernel: scsi: unknown device type 31

Device types are used to, well, tell the type of the attached device, e.g. disk, tape drive, media changer, CD-ROM, ...

A disk has device type 0 in standard SCSI, I don't know what 31 means. I thought that the max. value is 15, but it might have been expanded for iSCSI (I don't have time to check, sorry).


> Sep 11 13:28:57 backup002 kernel: scsi: host 11 channel 0 id 0 lun2340 has a LUN larger than allowed by the host adapter

Many host bus adapters and device drivers limit the addressable LUN range to 0..255. That looks like a presentation error on the NetApp to me.

[Maaz]

Thanks Uwe Zessin for help

> Many host bus adapters and device drivers limit the addressable LUN range > to 0..255.

hmm.. sorry didnt get this "addressable LUN range to 0..255" ? please explain.

>That looks like a presentation error on the NetApp to me.

I think you are right, because when we assign the LUN of 5 GB or above, we didnt receives the following message:
"kernel: scsi: host 11 channel 0 id 0 lun2340 has a LUN larger than allowed by the host adapter"
The above message only receives when we assign 1024 MB LUN.

[Uwe Zessin]

Think of it like a 'tree' structure.

Each SCSI adapter is one node. It might have one or more channels - on every channel is a 'bus'.

On each bus there are zero to many targets, numbered 0..n.

A target can have one or more LUNs. Depending on the type of SCSI protocol, the range of valid LUN addresses can be
- between 0..7 (parallel SCSI)
- 0..31 (some advanced SCSI)
- 0..255 (Peripherial Device LUN addressing, by many Fibre Channel components)
- 0..16384? (Volume Set Addressing, as used by HP-UX)

> kernel: scsi: host 11 channel 0 id 0 lun2340

Host 11 is the address of the iSCSI software adapter, channel 0, [target] id 0, LUN [adress] 2340

[skt_skt]

""kernel: scsi: host 11 channel 0 id 0 lun2340 has a LUN larger than allowed by the host adapter"

AFAIK, there isn't any limitation on the "smaller" LUN . Since the message is from the kernel i would recommend your iscsi configuration file (/etc/iscsi.conf or which ever is applicable in your case) or any adapter specific configuration file( i am not sure if any)

[Maaz]

Hi Thanks Santhosh Kumar Theyyan for help.

If you are asking for iSCSI configuration files/directories when the LUN was 1024 MB(that I deleted because its not working), so I cant provide you because I have deleted the LUN(1024 MB), destroy the configuration at iscsi-initiator(sles10).
But I am sending the configuration file(/etc/iscsi) from the same iscsi-initiator(sles10) machine, but this time LUN size is 7 GB.

[Maaz]

and forget to mention that when I assign a LUN size of 7 GB it works no issue, but when I assigned a LUN of 1024 MB, it cause the issue.

so the tarball attached in the previous post, is from a properly working iscsi-client(sles10)

[Uwe Zessin]

Interesting...

1TB is 2^31 = 2147483648(10) = 80000000(16) blocks

Is it possible that some component does work with signed integers??

[Maaz]

>1TB is 2^31 = 2147483648(10) = 80000000(16) blocks

>Is it possible that some component does work with signed integers??

I didnt get you Dear Uwe Zessin ? I mean how did you got the "1 TB" I mean from where you find 1 TB .. is it mention in the configuration file, I sent ?

[Uwe Zessin]

Maaz,

you wrote that there is a problem with a LUN size of 1024 MB.
1024MB = 1TB.

Can you create a 999MB or 1023MB LUN und see if access it still possible?

[Maaz]

Dear Uwe Zessin

>1024MB = 1TB.
1024 MB = 1 GB

>Can you create a 999MB or 1023MB LUN und see >if access it still possible?
I'll let you know

[Maaz]

after checking a lot many time I now found that issue/problem is not the LUN size, I assigned 1 GB, 2 Gb, 4 Gb and 7 and 10 GB luns to the sles10(iscsi-initiator), and got the same error

i.e
# lsscsi
[11:0:0:0] no dev NETAPP LUN 0.2 -

# tail -f /var/log/messages
ep 23 03:40:10 test kernel: scsi4 : iSCSI Initiator over TCP/IP
Sep 23 03:40:10 test kernel: scsi: unknown device type 31
Sep 23 03:40:10 test kernel: Vendor: NETAPP Model: LUN Rev: 0 .2
Sep 23 03:40:10 test kernel: Type: Unknown ANSI S CSI revision: 04
Sep 23 03:40:10 test kernel: 4:0:0:0: Attached scsi generic sg2 type 31
Sep 23 03:40:10 test kernel: scsi: host 4 channel 0 id 0 lun777 has a LUN larger than allowed by the host adapter
Sep 23 03:40:10 test iscsid: connection1:0 is operational now
Sep 23 03:40:16 test iscsid: iSCSI logger with pid=4562 started!
Sep 23 03:40:16 test iscsid: transport class version 2.0-724. iscsid version 2.0 -868
Sep 23 03:40:16 test iscsid: Can not bind IPC socket

I have attached the /etc/iscsi directory, from the machine(sles10 iscssi-inititator) which is facing/having the said/above problem

Regards

[Maaz]

from the same problem facing machine

# rcopen-iscsi restart
Closing all iSCSI connections: Logging out of session [sid: 1, target: iqn.1986- 03.com.ibm:sn.135025842, portal: 10.200.3.207,3260]
Logout of [sid: 1, target: iqn.1986-03.com.ibm:sn.135025842, portal: 10.200.3.20 7,3260]: successful
done
Stopping iSCSI initiator service: done
Starting iSCSI initiator service: done
Attempting discovery on target at 10.200.3.207: done
Setting up iSCSI targets: Logging in to [iface: default, target: iqn.1986-03.com .ibm:sn.135025842, portal: 10.200.3.207,3260]
Login to [iface: default, target: iqn.1986-03.com.ibm:sn.135025842, portal: 10.2 00.3.207,3260]: successful
done


test:/etc/iscsi # iscsiadm -m node
10.200.3.207:3260,2000 iqn.1986-03.com.ibm:sn.135025842

[JuergenGrabowski]

Hello,
i have the same error. Have you found a solution?