Re: open-iscsi issue

[Farhan Ahmed]

Thanks Mike

Here is the file attached

I noticed one more thing , If i disable eth2 I can ping the SAN array from
Linux via eth3/eth5/eth6 interfaces but I cant ping any server
eth3/eth5/eth6 interface from the equallogic array.

Regards,

Farhan




On Tue, Aug 2, 2011 at 1:52 PM, Mike Christie  wrote:

> On 07/29/2011 01:36 AM, Farhan Ahmed wrote:
> > I ran three commands and their tpcdump output is attached
> >
> > 1) iscsiadm -m discovery -t st -p 192.168.42.190 -I eth3 -d 8
> >
> > 2) iscsiadm -m discovery -t st -p 192.168.42.190 -I eth3 -d 8   // -l
> login
> >
> > 3) iscsiadm -m node -T
> > iqn.2001-05.com.equallogic:0-8a0906-f0a1ed402-1f40012bba54e28b-nfs02-bond
> -I
> > eth3 -l
> >
>
>
> You have to run tcpdump as
>
> tcpdump -i eth3 -w mytcpdumpfile
>
> then send me mytcpdumpfile
>

-- 
You received this message because you are subscribed to the Google Groups 
"open-iscsi" group.
To post to this group, send email to open-iscsi@googlegroups.com.
To unsubscribe from this group, send email to 
open-iscsi+unsubscr...@googlegroups.com.
For more options, visit this group at 
http://groups.google.com/group/open-iscsi?hl=en.



mytcpdumpfile
Description: Binary data

Re: open-iscsi issue

[Mike Christie]

Hi Shyam and Jim,

Do you know the answer to or can you redirect me to someone that can
answer the following:

The user below, Farhan Ahmed, is doing the following with a Equallogic
target.


On the initiator box he has multiple network interfaces on the same
subnet. He then has a iscsi iface for each one that is bound to the
iscsi target. So he has run

iscsiadm -m discovery -t st -p ip -I his_iface1


If he disables something in the network that affects the default route
then he can do:

ping -I net_interface $targets_ip


but when we try to login a normal session it will fail. See packet #97.

In this test he does not run something like route on the initiator box
to update the default route. Instead he is relying on the iscsi code's
interface binding to force the use of the interface.

We can see the iscsi login pdu get sent, but we do not see a iscsi
response. From what I can tell we just seem to see a tcp ack. Then later
in packet #104 we see the initiator give up on waiting for a login
response and disconnect the tcp/ip connection.


However, the weird thing is that before we try the normal session login
we see in packet #10 that a discovery session works ok in this test.

So command #1 works but command #3 does not and between that time he has
not done any changes to the iscsi config or network.

Is what Farhan is trying to do supported by EQL targets?

And is there a way to if the target is getting the iscsi login pdu
correctly?


On 08/01/2011 11:01 PM, Farhan Ahmed wrote:
> Thanks Mike
> 
> Here is the file attached
> 
> I noticed one more thing , If i disable eth2 I can ping the SAN array
> from Linux via eth3/eth5/eth6 interfaces but I cant ping any server
> eth3/eth5/eth6 interface from the equallogic array.
> 
> Regards,
> 
> Farhan
> 
> 
> 
> 
> On Tue, Aug 2, 2011 at 1:52 PM, Mike Christie  > wrote:
> 
> On 07/29/2011 01:36 AM, Farhan Ahmed wrote:
> > I ran three commands and their tpcdump output is attached
> >
> > 1) iscsiadm -m discovery -t st -p 192.168.42.190 -I eth3 -d 8
> >
> > 2) iscsiadm -m discovery -t st -p 192.168.42.190 -I eth3 -d 8   //
> -l login
> >
> > 3) iscsiadm -m node -T
> >
> iqn.2001-05.com.equallogic:0-8a0906-f0a1ed402-1f40012bba54e28b-nfs02-bond
> -I
> > eth3 -l
> >
> 
> 
> You have to run tcpdump as
> 
> tcpdump -i eth3 -w mytcpdumpfile
> 
> then send me mytcpdumpfile
> 
> 

-- 
You received this message because you are subscribed to the Google Groups 
"open-iscsi" group.
To post to this group, send email to open-iscsi@googlegroups.com.
To unsubscribe from this group, send email to 
open-iscsi+unsubscr...@googlegroups.com.
For more options, visit this group at 
http://groups.google.com/group/open-iscsi?hl=en.



mytcpdumpfile
Description: Binary data

Re: [PATCH 1/7] Fix dcb_app.c compile error with old kernels

[Mike Christie]

Merged all these patches. They should show up on the kernel.org git tree
in a couple of hours.

On 07/20/2011 08:06 AM, Jim Ramsay wrote:
> Linux kernels 2.6.28 and earlier do not have RTM_GETDCB or RTM_SETDCB
> defined in their rtnetlink.h.  Defining them in dcb_app.c should be safe
> since they have not changed their numerical value since they were
> introduced in 2.6.29.
> 

-- 
You received this message because you are subscribed to the Google Groups 
"open-iscsi" group.
To post to this group, send email to open-iscsi@googlegroups.com.
To unsubscribe from this group, send email to 
open-iscsi+unsubscr...@googlegroups.com.
For more options, visit this group at 
http://groups.google.com/group/open-iscsi?hl=en.

Re: iscsi_sysfs_read_iface

[Mike Christie]

On 07/26/2011 01:38 AM, Hannes Reinecke wrote:
> diff --git a/usr/iscsi_sysfs.c b/usr/iscsi_sysfs.c
> index e82fe80..145816d 100644
> --- a/usr/iscsi_sysfs.c
> +++ b/usr/iscsi_sysfs.c
> @@ -489,7 +489,7 @@ static int iscsi_sysfs_read_iface(struct iface_rec
> *iface, int host_no,
> if (ret) {
> ret = sysfs_get_str(id, ISCSI_HOST_SUBSYS, "initiatorname",
> iface->iname, sizeof(iface->iname));
> -   if (ret)
> +   if (ret) {
> /*
>  * default iname is picked up later from
>  * initiatorname.iscsi if software/partial-offload.
> @@ -499,6 +499,8 @@ static int iscsi_sysfs_read_iface(struct iface_rec
> *iface, int host_no,
>  */
> log_debug(7, "Could not read initiatorname for "
>   "host%d\n", host_no);
> +   ret = 0;
> +   }
> }
> 

Added in commit 4cc22470d5a1391108f329842f63ff3c31ccd288. Thanks.

-- 
You received this message because you are subscribed to the Google Groups 
"open-iscsi" group.
To post to this group, send email to open-iscsi@googlegroups.com.
To unsubscribe from this group, send email to 
open-iscsi+unsubscr...@googlegroups.com.
For more options, visit this group at 
http://groups.google.com/group/open-iscsi?hl=en.