redhat-cluster-suite fails to build from source in maverick

http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=502401, already fixed by a debian commit. Shall work on getting it in Ubuntu's redhat-cluster package

Please ignore the previous comment. The debian error is related to gfs, but different. the debian/changelog says this is a gfs build error.

/usr/include/linux/gfs2_ondisk.h is a reflection of linux-2.6/include/linux/gfs2_ondisk.h. Between maverick and lucid, this .h in the linux-2.6 source code has changed; the changes involve removal of some structure members.

1) These changes/removals are not yet reflected in /usr/include/linux/gfs2_ondisk.h.
2) The reason why this compilation is failing is that these changes/removals are also not reflected in the gfs2/libgfs2(/ondisk.c) code. However owing to the changes/removal of members from structures, the compiler cannot find the member in the structure definition and complains.
The correct solution for this would be to selectively change the code that uses this. However a better solution is to merge the maverick redhat-cluster with debians latest redhat-cluster.

Surbhi,

As part of the blueprint https://blueprints.launchpad.net/ubuntu/+spec/server-maverick-clusterstack we deal with the merge/sync of cluster related packages, which include redhat-cluster.

Now, we've not yet tested the new redhat-cluster package in debian, which might indeed fix the FTBFS. However, AFAIK, the new redhat-cluster package in Debian has changes than involve the use of dependencies that in Ubuntu are currently in Universe and not main, and if we do merge it now, it will FTBFS.

The MIRs are going to be filed soon to include these cluster related packages into Main. Once this is done, we'll deal with RHCS.

Thank you.

The packages in question that would need to be promoted appear to be cluster-glue and pacemaker. There's no problem with syncing or merging redhat-cluster first and letting it fail to build while the MIRs are processed; if nothing else, packages are not typically promoted to main until there is a reason to do so in the archive, so it's actually best to have the new redhat-cluster in place first so that it shows up on the archive administrators' report of packages that need to be promoted.

Better solution then. I'll take a look to rhcs, run some tests and then file the sync/merge. :)

This has already been fixed upstream.

Yes, that's why we are gonna sync (or merge if necessary) from debian. However, pacemaker is not yet in Main which will make redhat-cluster to FTBFS until pacemaker gets promoted to main!

Attaching debdiff.

This bug was fixed in the package redhat-cluster - 3.0.12-2ubuntu1

---------------
redhat-cluster (3.0.12-2ubuntu1) maverick; urgency=low

  * Resync from debian unstable. Fixes FTBFS (LP: #600984)
  * debian/gfs2-tools.links: Fix symlinks.
  * debian/control: Build-Depends on libltdl7-dev.

redhat-cluster (3.0.12-2) unstable; urgency=low

  * Upload to unstable

redhat-cluster (3.0.12-1) experimental; urgency=low

  * [a45231b] New patch 0005-Remove-bashisms-from-resource-scripts.patch
    Remove bashisms from resource scripts (Closes: #581137)
  * [3df4d28] Imported Upstream version 3.0.12
  * [25ffc05] Install rgmanager manpage
  * [7e5f87f] Update jp translation (Closes: #582722) - thanks to Hideki
    Yamane

redhat-cluster (3.0.11-1) experimental; urgency=low

  * [9c55372] reformat ja.po
  * [ca106f6] Imported Upstream version 3.0.11
  * [03336a5] Grab libdlmcontrol.so.3.1
  * [c14c894] Enable pacemaker support - thanks a lot to Jörg Bachmann for
    testing
  * [287d065] New patch 0004-Depend-on-remote_fs.patch Depend on $remote_fs
  * [54bfc48] Bump (build-)dep on corosync.

redhat-cluster (3.0.9-2) experimental; urgency=low

  * Upload to unstbale
  * [fb1da7e] Unignore po files ignored by upstreams .gitignore file
  * [6845591] Add new vi translation (Closes: #576145) - thanks to Clytie
    Siddall
  * [aa6bf55] Add jp debconf translation (Closes: #554841) - thanks to Hideki
    Yamane

redhat-cluster (3.0.9-1) experimental; urgency=low

  * [7c1a75c] Imported Upstream version 3.0.7
  * [9381ece] Drop 0003-Do-nothing-if-cman_tool-isn-t-there.patch fixed
    upstream
  * [8ed7908] Bump (build-)dep on corosync
  * [5b162fa] Enable gfs since it defaults to off now upstream.
  * [1d4ad8f] Imported Upstream version 3.0.9
  * [09c214b] Version dependencies of meta package
  * [4352b38] Redo patches
  * [61487cc] New patch 0003-Don-t-rely-on-chkconfig-and-check-for-
    network-manage.patch Don't rely on chkconfig and check for network-
    manager

redhat-cluster (3.0.6-5) unstable; urgency=low

  * [ebda28c] Don't ship drbd resource script (Closes: #566095)

redhat-cluster (3.0.6-4) unstable; urgency=low

  * [9725a65] Only install schema (Closes: #565382)

redhat-cluster (3.0.6-3) unstable; urgency=low

  * Upload to unstable
  * [19bd29c] Depend on cman (Closes: #563787)
  * [2a019fa] Genereate proper depends for lcsos
  * [e276e1d] Add schema

redhat-cluster (3.0.6-2) experimental; urgency=low

  * [e16890c] Update de.po (Closes: #562707) - thanks to Helge Kreutzmann
  * [e1b336d] Update pt debconf translation (Closes: #563015) - thanks to
    Américo Monteiro
  * [e01aa76] Update ru debconf translation (Closes: #563345) - thanks to Yuri
    Kozlov
  * [f959fd1] Update sv debconf translation (Closes: #562928) - thanks to
    Martin Bagge
  * [bbb4b4a] Run debconf-updatepo

redhat-cluster (3.0.6-1) experimental; urgency=low

  * [b3b823a] Imported Upstream version 3.0.6
  * [3e69621] run debconf-updatepo during clean
  * [cb83708] build-depend on po-debconf
  * [b8b59d7] remove one more trailing whitespace
  * [ccd9a39] Depend on libxml2-utils for ccs_config_validate
  * [ff09490] redo patches
  * [08a23fa] Do nothing if cman_tool is...