Andy Grover | 7 Oct 20:38 2014
Picon

[PATCHv2 00/13] Warning cleanups

Changes in v2:
* Fix mac_address size issue
* Fix issues with unused but set variables instead of disabling the
  warning

Thanks -- Andy

Andy Grover (13):
  Fix build warnings for unused variables
  Fix warning about possibly-uninitialized variable
  Fix bad sizeof in memset
  Fix missing header
  iscsiuio: Fix warning about non-matching types
  iscsiuio: Fix strict-aliasing warning with struct mac_address
  iscsiuio: Resolve strict aliasing issue in iscsiuio/src/unix/nic.c
  iscsiuio: Fix aliasing issue with IPV6_IS_ADDR_UNSPECIFIED
  iscsiuio: Use attribute(unused) for variables that are unused but
    needed
  iscsiuio: Use attribute(unused) for *icmpv6_hdr
  iscsiuio: Change nic_disable to return void
  iscsiuio: Remove set but unused variables
  iscsiuio: Check return value from nic_queue_tx_packet

 iscsiuio/src/apps/dhcpc/dhcpv6.c |  6 ++----
 iscsiuio/src/uip/ipv6.c          |  2 +-
 iscsiuio/src/uip/ipv6.h          | 39 ++++++++++++---------------------------
 iscsiuio/src/uip/pt.h            |  2 +-
 iscsiuio/src/uip/uip.c           |  2 +-
 iscsiuio/src/uip/uip_arp.c       |  2 --
 iscsiuio/src/unix/libs/bnx2x.c   |  2 +-
(Continue reading)

Jamie Clark | 24 Sep 10:01 2014

Occasional data digest errors

On machines running Ubuntu 12.04 open-iscsi initiator we have noticed intermittent "pdu ... rejected due to DataDigest error.  (data digests configured and negotiated ok on initiator and target).

After extensive wiresharking (the problem only affects something like 1 in 100000 write PDUs) I captured a failed write from the initiator and found that the CRC32C value following the payload was incorrect, while surrounding writes were OK.

The target then rejects this bad pdu and follows that with an RST on the TCP connection. This is expected behaviour. This seems to repeat about every 10 minutes or so.

The writes are all 4k blocks originating from iscsi-backed libvirt managed KVM virtual machines. The particular failure I captured appeared to be syslog data. The first 100 bytes or so of the 4k data was a syslog entry, the remainder was zeroed out.

Just wondering if digest calculation still has the potential failure modes described back here https://groups.google.com/d/msg/open-iscsi/OCYS5wUIYhc/Gno7edWCyH0J

The stack looks something like this:



      +-------------------+
      | rsyslog           |
      |    +              |
      |    | write()      |
      |    v              |
      | /var/log/messages |
      |    +              |
      |    |              |
      |    v              |
      | /dev/vdb          |
      +-------------------+
           +            qemu/kvm VM virtio format=raw cache=none
           |
           v
      /dev/mapper/xxx  (dm-multipath)
           +
           |
           v
      /dev/{sdx,sdy}   (iscsi LUN)
           +
           |
           v
         eth2          (packet capture)

--
You received this message because you are subscribed to the Google Groups "open-iscsi" group.
To unsubscribe from this group and stop receiving emails from it, send an email to open-iscsi+unsubscribe-/JYPxA39Uh5TLH3MbocFFw@public.gmane.org.
To post to this group, send email to open-iscsi-/JYPxA39Uh5TLH3MbocFFw@public.gmane.org.
Visit this group at http://groups.google.com/group/open-iscsi.
For more options, visit https://groups.google.com/d/optout.
Lee Duncan | 26 Sep 18:44 2014
Picon

Re: PATCH: two patches that update network state access for open-iscsi


On Sep 25, 2014, at 12:47 AM, open-iscsi@... wrote: 
> Deviants will be sacrificed to ensure group solidarity.
> Mike Christie <michaelc@...>: Sep 24 11:58AM -0500 
> 
> On 09/24/2014 10:59 AM, Lee Duncan wrote:
>  
> > 0002-Fixup-IPv6-iBFT-interface-description.patch -- This adds interface
> > prefix length to handle some kernels that have an issue with respect to
> > iBFT, as well as adding the code in open-iscsi to use this value, if present
>  
> What is the prefix? Could you give me an example? We are reading this
> from the kernel right?
>  
> + sysfs_get_int(id, subsys, "prefix-len", &context->prefix);
>  
> I do not see that in the upstream kernel. Is it in one of Konrad's trees?

Hannes added this to the SUSE Kernel July of this year, and submitted it to LKML.

The issue is that IPv6 does not use a netmask, but instead exclusively uses a prefix length. I didn't realize
that this had not made it upstream yet.

>  
> Also, it seems this second patch does several things. Could you separate
> out the coding style fixes (go ahead and include spelling changes in
> there) and coding style changes that you seem to prefer, the
> context->origin changes (what was the reason for them), and the flag
> changes?

Yes, I will resubmit.

>  
> And for the flag changes, did you test this patch with be2iscsi boot? It
> looks like it sets the flags to 2, so it will fail right?
> 

I believe so, but I will make sure that is the case before I resubmit.

-- 
Lee-Man Duncan

"Beer is proof that God loves us and wants us to be happy." -- Ben Franklin

--

-- 
You received this message because you are subscribed to the Google Groups "open-iscsi" group.
To unsubscribe from this group and stop receiving emails from it, send an email to open-iscsi+unsubscribe@...
To post to this group, send email to open-iscsi@...
Visit this group at http://groups.google.com/group/open-iscsi.
For more options, visit https://groups.google.com/d/optout.

Lee Duncan | 24 Sep 17:59 2014
Picon

PATCH: two patches that update network state access for open-iscsi

Hi Mike:

Here are a couple of patches that add more /sys access to open-iscsi, both created by Hannes and part of our tree.

0001-Allow-modifications-for-iface.gateway-and-iface.subn.patch -- This adds interface gateway and subnet mask access in /sys, and

0002-Fixup-IPv6-iBFT-interface-description.patch -- This adds interface prefix length to handle some kernels that have an issue with respect to iBFT, as well as adding the code in open-iscsi to use this value, if present

The patches are attached.
v>
-- 
You received this message because you are subscribed to the Google Groups "open-iscsi" group.
To unsubscribe from this group and stop receiving emails from it, send an email to open-iscsi+unsubscribe@...
To post to this group, send email to open-iscsi@...
Visit this group at http://groups.google.com/group/open-iscsi.
For more options, visit https://groups.google.com/d/optout.


-- 
The Lee-Man

Deviants will be sacrificed to ensure group solidarity.



khagsted | 17 Sep 20:33 2014
Picon

open-iSCSI with BluRay drive use sr driver scsi-1 instead of scsi3-mmc.

Dear List

I have a server with a BluRay drive that is shared as an iSCSI target, with either SCST or LIO (I have tried both too see if the error was on the target implementation). It is possible for me to connect to this drive both from Linux using open-iSCSI and from windows using the MS-iSCSI-initiator. On windows I can use the drive like it was a local drive, meaning that I can play all type of content from it including encrypted DVDs and I would also be able to change the region settings on the drive if I felt for it. However on Linux with open-iSCSI I'm not able to play encrypted DVDs, unless a MS-Windows computer is also connected to the drive. I am also not able to retrieve the region settings of the drive with the regionset utility. After much poking around I found some differences between the server system and the initiator system when adding then binding the drive to the sr module.

For the  server system, which can play encrypted DVDs the log entries are:
[    6.661163] scsi 2:0:0:0: CD-ROM            Optiarc  BD ROM BC-5600S  1.82 PQ: 0 ANSI: 5
[    6.675635] sr0: scsi3-mmc drive: 16x/16x writer dvd-ram cd/rw xa/form2 cdda tray
[    6.675639] cdrom: Uniform CD-ROM driver Revision: 3.20
[    6.675801] sr 2:0:0:0: Attached scsi CD-ROM sr0
[    6.675910] sr 2:0:0:0: Attached scsi generic sg2 type 5

For the initiator system, which cannot play encrypted DVDs the log entries are:
[   36.650119] scsi 9:0:0:0: CD-ROM            Optiarc  BD ROM BC-5600S  1.82 PQ: 0 ANSI: 5
[   36.663525] sr0: scsi-1 drive
[   36.663530] cdrom: Uniform CD-ROM driver Revision: 3.20
[   36.663806] sr 9:0:0:0: Attached scsi CD-ROM sr0
[   36.664068] sr 9:0:0:0: Attached scsi generic sg4 type 5

Please note here the difference in the second entry on the two systems, one says "sr0: scsi3-mmc drive" and the other "sr0: scsi-1 drive". By looking through the linux kernel code I found out that the module falls back to scsi-1 if it cannot read the capabilities mode page of the attached drive. Please refer to "drivers/scsi/sr.c" from around line 873 in the Linux source code. This, for me, indicates that open-iSCSI does not set-up the access to the pass-throughed device correctly.

It should be noted that I am running on Ubuntu 14.04, with libdvdread4 and libdvdcss2 installed correctly.

I hope someone can help me fix this problem as I would like to switch completely to Linux.

Please fell free to ask for further info, and best regards Kristian Hagsted Rasmussen

--
You received this message because you are subscribed to the Google Groups "open-iscsi" group.
To unsubscribe from this group and stop receiving emails from it, send an email to open-iscsi+unsubscribe-/JYPxA39Uh5TLH3MbocFFw@public.gmane.org.
To post to this group, send email to open-iscsi-/JYPxA39Uh5TLH3MbocFFw@public.gmane.org.
Visit this group at http://groups.google.com/group/open-iscsi.
For more options, visit https://groups.google.com/d/optout.
Rob Roschewsk | 15 Sep 17:06 2014
Picon

iscsistart Ubuntu server 14.01.1 LTS / HP proliant DL380p GEN 8

OK ... so I've made a lot of progress since my last post about VLAN id .... discovered I should be using ISCSI_AUTO and ibft .... However ... I'm still getting grief ....

from the init script local-top/iscsi, "iscsistart -b" is failing with the message:
Could not get boot entry.

I have hacked at the iscsi script and pared it down to the bare essentials:
echo "loading ibft module"
modprobe iscsi_ibft
echo "load ibft module"
echo "sleep 5"
sleep 5
echo "wake up"
echo "running iscsistart"
/sbin/iscsistart -b
echo "ended iscsistart"
echo "running udevadm settle"
udevadm settle
echo "ended udevadm settle"
echo "ended iscsi script"

.... no joy ....

but here is the kicker .... 

when the initscript fails to find root and drops me to the initramfs shell .... I run "iscsistart -b" at the command line it all works fine! .... configures the hba and connects to the lun .... on both interfaces .. exit out of the shell ... root gets mounted and all is well.

The box is using emulex lan/iscsi adapters:
03:00.2 Mass storage controller: Emulex Corporation OneConnect 10Gb iSCSI Initiator (be3) (rev 01)
03:00.3 Mass storage controller: Emulex Corporation OneConnect 10Gb iSCSI Initiator (be3) (rev 01)


I'm stumped ... is it a timing thing??? Any and all help appreciated !!!

Thanks!

--> Rob





--
You received this message because you are subscribed to the Google Groups "open-iscsi" group.
To unsubscribe from this group and stop receiving emails from it, send an email to open-iscsi+unsubscribe-/JYPxA39Uh5TLH3MbocFFw@public.gmane.org.
To post to this group, send email to open-iscsi-/JYPxA39Uh5TLH3MbocFFw@public.gmane.org.
Visit this group at http://groups.google.com/group/open-iscsi.
For more options, visit https://groups.google.com/d/optout.
Kelley, Jared | 11 Sep 20:38 2014

amazonAWS/VTL with open-iscsi on EC2 /dev/by-path/ not present

I’m trying to setup a VTL on some AWS EC2 instances and have done so with success however the scsi devices (virtual tape drives)
show up as connected to the client but there is nothing listed under /dev/st* nor is there any /dev/by-path directory on the client.

Has anyone experienced this and what might be the issue?

Thanks in advance

Jk

--
You received this message because you are subscribed to the Google Groups "open-iscsi" group.
To unsubscribe from this group and stop receiving emails from it, send an email to open-iscsi+unsubscribe-/JYPxA39Uh5TLH3MbocFFw@public.gmane.org.
To post to this group, send email to open-iscsi-/JYPxA39Uh5TLH3MbocFFw@public.gmane.org.
Visit this group at http://groups.google.com/group/open-iscsi.
For more options, visit https://groups.google.com/d/optout.

Tejas vaykole | 11 Sep 12:22 2014
Picon

SHA-1 hashing Algorithm for CHAP

Hello,

I am trying out with the open-iscsi initiator.I see that the initiator uses MD5 algorithm for CHAP.
I need help in configuring the initiator to use SHA-1 hashing Algorithm for CHAP.

Thanks.
Tejas

--
You received this message because you are subscribed to the Google Groups "open-iscsi" group.
To unsubscribe from this group and stop receiving emails from it, send an email to open-iscsi+unsubscribe-/JYPxA39Uh5TLH3MbocFFw@public.gmane.org.
To post to this group, send email to open-iscsi-/JYPxA39Uh5TLH3MbocFFw@public.gmane.org.
Visit this group at http://groups.google.com/group/open-iscsi.
For more options, visit https://groups.google.com/d/optout.
Rob Roschewsk | 11 Sep 17:43 2014
Picon

iscsistart VLAN ID ??

Hi all ... I've searched but have found no answers .... maybe my google-juice is low ...

I need to be able to pass a VLAN ID and enable VLAN TAGGING in iscsistart 

I'm using an EMULEX 10G NETWORK / ISCSI HBA and our iscsi traffic is on a specific vlan 

Emulex Corporation OneConnect 10Gb NIC (be3) (rev 01)

I am able to connect to the device from the command line using iscsiadm .... but during the boot process I want to mount the iscsi lun as root.

I see nothing in the scripts / configs / man pages about VLAN support ....

FYI ... trying to do this on Ubunti Server 14.04.1

Thanks ...

--> Rob

--
You received this message because you are subscribed to the Google Groups "open-iscsi" group.
To unsubscribe from this group and stop receiving emails from it, send an email to open-iscsi+unsubscribe-/JYPxA39Uh5TLH3MbocFFw@public.gmane.org.
To post to this group, send email to open-iscsi-/JYPxA39Uh5TLH3MbocFFw@public.gmane.org.
Visit this group at http://groups.google.com/group/open-iscsi.
For more options, visit https://groups.google.com/d/optout.
The Lee-Man | 12 Sep 23:56 2014
Picon

A pair of patches to add utility scripts

Here are a pair of patches that each add a new utility script.

0001-Added-new-utility-script-to-generate-initiator-name.patch -- Adds
a script in utils calls iscsi-gen-initiatorname, which we use in our distribution
to create an initial iSCSI initiator name, but which can also be useful
to update/change the name.

0002-Added-new-util-script-to-aid-in-CNA-setup.patch -- Adds
a script in utils called iscsi_offload, which helps set up iSCSI
offload using open-iscsi, including module loading and interface
setup.

Both scripts were created by Hannes and are now used in SLES and openSUSE,
but they may be of use to others.
--
Lee Duncan
SUSE Labs

--
You received this message because you are subscribed to the Google Groups "open-iscsi" group.
To unsubscribe from this group and stop receiving emails from it, send an email to open-iscsi+unsubscribe-/JYPxA39Uh5TLH3MbocFFw@public.gmane.org.
To post to this group, send email to open-iscsi-/JYPxA39Uh5TLH3MbocFFw@public.gmane.org.
Visit this group at http://groups.google.com/group/open-iscsi.
For more options, visit https://groups.google.com/d/optout.
mike | 6 Sep 12:06 2014

Delayed-OPT-ISCSI start

https://github.com/mikechristie/open-iscsi/issues/7#issuecomment-54693881


my Linux distro now rams it down new kernels. and patches.

the service loader wont FAIL...... or disable.

<at> Login screan or as X-loads , in backgound load service, if Kernel module fails
(sys-block / iscsitarget builds for newer kernels pending)
DBUS message. if service fails dbus message to root / username...

since ISCSI wont fail , my laptop is a nice brick with their stock kernel.....
using genkerel-next to make one for systemD to load GRRRRRR. so removing it not good.

but since the service wont gracefully fail , my laptop infinitely loops on and on. trying to load.

prior to that i could remove open-ISCSI and the km Package and go for a fresh install. (Problem solved)

Try N times load kernel module
initiator not found sorry cant load ISCSI,


ok np , Sabayon is a blend of/variant Gentoo , however it's always bleeding edge.
I have no qualms using ISCSI , however enabling at boot time of late has been the kiss of death to boot up. letting the boot goto Login: and then in background load it then as a delayed start up and standard background service.

However if you make SystemD services files that on boot , DO N times , fail..... once login: is up I can fix and or manually run a service re-start.
or a delayed start-up service....  I will gleefully test them.
but i'll post out to the group.....


esle load KM , if ISCSI is bugged IE build from binary isn't working and fails ( ie bugs) load N times
Fail send alert.

since my root is a physical drive and on my laptop I could care less least for Boot-time load.
now if i build up a Freenas having it run to keep my backups and repos (ie github) etc , oh yeah iscsi quite good for that , and or client shares.

ISUE="-Boot-start +Delayed-start"

Sabayon/Gentoo isn't the only distro suffering, IE Redhat were Sabayon ports things to the Gentoo land etc, is also affected...

if your a mobile user with quirky hardware and a bleeding edge distro , shit happens like Nvida drivers... download the *.run and do install works ever time. however since the open-iscis breaks and when it breaks it really breaks it's now forced , were as previously i could remove it temporally .

--
You received this message because you are subscribed to the Google Groups "open-iscsi" group.
To unsubscribe from this group and stop receiving emails from it, send an email to open-iscsi+unsubscribe-/JYPxA39Uh5TLH3MbocFFw@public.gmane.org.
To post to this group, send email to open-iscsi-/JYPxA39Uh5TLH3MbocFFw@public.gmane.org.
Visit this group at http://groups.google.com/group/open-iscsi.
For more options, visit https://groups.google.com/d/optout.

Gmane