kickstart iscsi option broken

Do you have the iscsi initiator setup at the SAN? Sorry I am not expert on kickseed but I know quite a bit about SANs

@Taha:
The SAN setup is not the problem.

If I do a manual Ubuntu installation and enter the same portal IP-address, and select the same IQN out of the list, it installs fine.
The problem is only when doing automated installation through kickstart.

Hi,

I have the same problem.
I have tried using both KickSeed & raw preseed directives and get the same issue with both.
I am wondering if the blank username and password are the problem. I will try adding CHAP auth to the LUN and see if it makes any difference.
Irritating!

Update: Just tried with CHAP and it still hung.
I'm pretty sure this isn't a bug in KickSeed but it might be a Partman-ISCSI bug..

As I emailed to Colin earlier today this is because kickseed is omitting the datatype part of the kickseed config when performing the conversion from kickstart -> preseed.

eg

ks.cfg:

iscsi --ipaddr=xxx.xxx.xxx.xxx --target=iqn.1992-08.com.xxxxx:xxxx --port=3260

should be translated to:

d-i partman-iscsi/login/address string xxx.xxx.xxx.xxx:3260

but the "string" part is missing so preseed ignores it.

you can check this by dropping to a shell in the installer (alt+f2) and running

  debconf-dumpdb|grep iscsi

which will show you which preseed commands succesfully made it into the database. if address shows as (null) then it didnt work!

This bug was fixed in the package partman-iscsi - 17

---------------
partman-iscsi (17) oneiric; urgency=low

  * Don't fail if debconf questions are preseeded (LP: #810068).
 -- Colin Watson <email address hidden> Wed, 20 Jul 2011 17:04:26 +0100

This bug was fixed in the package kickseed - 0.56ubuntu2

---------------
kickseed (0.56ubuntu2) oneiric; urgency=low

  * Fix iSCSI ks_preseed calls to include a type field (LP: #810068).
 -- Colin Watson <email address hidden> Wed, 20 Jul 2011 17:11:06 +0100

I should stress that I haven't actually tested this pair of fixes - they're based on the logs Tim sent me - and would very much appreciate real-world testing. I expect that the resulting set of fixes will need to go to Lucid/Maverick/Natty, since I think this code has been broken all the way along.

Hello Max, or anyone else affected,

Accepted kickseed into lucid-proposed, the package will build now and be available in a few hours. Please test and give feedback here. See https://wiki.ubuntu.com/Testing/EnableProposed for documentation how to enable and use -proposed. Thank you in advance!

Hello Max, or anyone else affected,

Accepted partman-iscsi into lucid-proposed, the package will build now and be available in a few hours. Please test and give feedback here. See https://wiki.ubuntu.com/Testing/EnableProposed for documentation how to enable and use -proposed. Thank you in advance!

Hello Max, or anyone else affected,

Accepted kickseed into natty-proposed. The package will build now and be available in a few hours. Please test and give feedback here. See https://wiki.ubuntu.com/Testing/EnableProposed for documentation how to enable and use -proposed. Thank you in advance!

Hello Max, or anyone else affected,

Accepted partman-iscsi into natty-proposed. The package will build now and be available in a few hours. Please test and give feedback here. See https://wiki.ubuntu.com/Testing/EnableProposed for documentation how to enable and use -proposed. Thank you in advance!

Is there really much point in fixing this in maverick now? It has lived most of its live with this bug, and will go EOL in two months.

Can anyone please try the current lucid dailies to verify that this is fixed? http://cdimage.ubuntu.com/lucid/

We need to do this in the next few days to get this fix into 10.04.4.

I've done some testing on this SRU for Lucid in preparation for 10.04.4.

1) Verified that without -proposed I get prompted as detailed in the original bug report

2) With -proposed enabled (and using the proposed mini.iso with cobbler) I was able to successfully install to an iSCSI target

I did however get prompted for the username for the iSCSI initiator despite the fact that it does not require authentication and I had to restrict the iSCSI initiator to presenting only one device to the client as the --target parameter is not supported in kickseed.

See the attached kickstart file for details.

'preseed partman-iscsi/login/incoming_username string' in a Kickstart file should work around this, but I think this is a separate bug (I do agree it's a bug).

This now seems sufficiently verified to warrant copying to -updates. It is not a regression, and although it's not perfect yet it is definitively an improvement.

This bug was fixed in the package kickseed - 0.54ubuntu1.10.04.2

---------------
kickseed (0.54ubuntu1.10.04.2) lucid-proposed; urgency=low

  * Fix iSCSI ks_preseed calls to include a type field (LP: #810068).

kickseed (0.54ubuntu1.10.04.1) lucid-proposed; urgency=low

  * Preseed partman-lvm/confirm_overwrite as well as partman-lvm/confirm.
  * Preseed partman-lvm/device_remove_lvm when confirming logvol results,
    since Kickstart doesn't have a separate control with a one-to-one
    correspondence to this (LP: #708548).
 -- Colin Watson <email address hidden> Fri, 20 Jan 2012 16:51:53 +0000

This bug was fixed in the package partman-iscsi - 14.1

---------------
partman-iscsi (14.1) lucid-proposed; urgency=low

  * Don't fail if debconf questions are preseeded (LP: #810068).
 -- Colin Watson <email address hidden> Fri, 20 Jan 2012 16:44:33 +0000

eol reached https://wiki.ubuntu.com/Releases