[Bug 1833586] Re: open-iscsi unconditionally requires ib_iser module which is not built on all Ubuntu kernels

[Christian Ehrhardt ]

Glad to find this discussed and solved already when waking up - thanks
Ryan and Steve!

To complete the discussions I'm adding a little reference as on IRC it seemed this question was left open:
"[20:40] <vorlon> rharper: right, I also fundamentally question why the open-iscsi package should hard-code the modules for loading if there's additional configuration required (not necessarily extra packages, but at least non-default config options that should maybe be detected to trigger on-demand module loading)"

I think the history on this is [1] which was converting an old
unconditional load (that would do || true, and therefore never make it
fail/bad) into something else which instead of the old code had the
negative effect that Steve reported here initially.

[1]: https://salsa.debian.org/linux-blocks-team/open-
iscsi/commit/592d3d1f6

-- 
You received this bug notification because you are a member of Ubuntu
Foundations Bugs, which is subscribed to open-iscsi in Ubuntu.
https://bugs.launchpad.net/bugs/1833586

Title:
  open-iscsi unconditionally requires ib_iser module which is not built
  on all Ubuntu kernels

Status in open-iscsi package in Ubuntu:
  Fix Released

Bug description:
  Currently, open-iscsi ships a /lib/modules-load.d/open-iscsi.conf
  which hard-codes the iscsi_tcp and ib_iser modules.  However, we ship
  open-iscsi as part of the standard ubuntu-server seed, but not all
  images that derive from this seed have kernels that include the
  ib_iser module.  In particular, we've noticed that the Raspberry Pi
  kernel does not.  This is noticeable because if a module doesn't
  exist, the systemd-modules-load service fails, and leaves the system
  boot in a degraded mode per systemd.

  And unfortunately ib_iser has no modaliases, so it's not as if this is
  something that can be switched to udev autoloading.

  Options for resolving this:

   - get the kernel team to set CONFIG_INFINIBAND_ISER=m on all kernels.  But this would increase kernel memory consumption unconditionally on possibly memory-constrained arm systems, for no real value.  (A spot check on another arch shows roughly 750K; not huge but not nothing)
   - drop ib_iser from debian/open-iscsi.kmod (why does use of iscsi dictate loading a particular infiniband module? there's nothing in the debian/changelog that explains)
   - vary the contents of /lib/modules-load.d/open-iscsi.conf (only sensible if we believe that this module would never be useful on any armhf)
   - remove open-iscsi from the server seed; historically, this is blocked by older versions of MAAS depending on mounting ephemeral images via iscsi, but perhaps the depends on open-iscsi could be made architecture-dependent, if only armhf is affected by this problem and MAAS doesn't support any armhf targets?

  This needs input from the Server Team.

To manage notifications about this bug go to:
https://bugs.launchpad.net/ubuntu/+source/open-iscsi/+bug/1833586/+subscriptions