Branches for Natty

Name Status Last Modified Last Commit
lp:ubuntu/natty/nfs-utils bug 2 Mature 2011-04-01 11:22:19 UTC
37. Search kerberos libs in multiarch loc...

Author: Matthias Klose
Revision Date: 2011-04-01 11:22:19 UTC

Search kerberos libs in multiarch locations.

lp:ubuntu/natty-proposed/nfs-utils bug 2 Mature 2011-08-17 12:54:40 UTC
38. * debian/nfs-common.idmapd.upstart: d...

Author: Steve Langasek
Revision Date: 2011-07-17 22:21:57 UTC

* debian/nfs-common.idmapd.upstart: don't use a script unnecessarily for
  our job when we can exec directly - making the job more resilient in
  the face of races with /usr being mounted. LP: #811823.
* Drop rpc_pipefs.conf; this has gotten far more complicated than it
  should be, just do the mount in-line in each of the gssd and idmapd
  jobs.

lp:ubuntu/natty-updates/nfs-utils 2 Mature 2011-10-20 17:25:22 UTC
38. * debian/nfs-common.idmapd.upstart: d...

Author: Steve Langasek
Revision Date: 2011-07-17 22:21:57 UTC

* debian/nfs-common.idmapd.upstart: don't use a script unnecessarily for
  our job when we can exec directly - making the job more resilient in
  the face of races with /usr being mounted. LP: #811823.
* Drop rpc_pipefs.conf; this has gotten far more complicated than it
  should be, just do the mount in-line in each of the gssd and idmapd
  jobs.

lp:~clint-fewbar/ubuntu/natty/nfs-utils/wait-for-local-filesystems bug(Has a merge proposal) 1 Development 2011-01-11 16:28:46 UTC
38. removing unnecessary debugging sleep ...

Author: Clint Byrum
Revision Date: 2011-01-11 16:28:46 UTC

removing unnecessary debugging sleep delay post-start

lp:~ken-vandine/ubuntu/natty/nfs-utils/1.2.2-4ubuntu1 bug 1 Development 2010-12-07 18:17:34 UTC
23. * Merge from debian unstable (LP: #68...

Author: Ken VanDine
Revision Date: 2010-12-07 18:16:09 UTC

* Merge from debian unstable (LP: #685860), remaining changes:
  - debian/nfs-common.{statd,gssd,idmapd,rpc_pipefs}.upstart,
    debian/control, debian/nfs-common.{preinst,postinst,prerm,postrm},
    debian/rules: drop nfs-common init script in favor of upstart jobs, and
    build-depend on debhelper (>= 7.3.15ubuntu3) for correct upstart init
    handling.
  - debian/control:
    + depend on the upstart-using version of portmap, 6.0-10ubuntu1; and
      drop the alternative depends on rpcbind, which hasn't been converted.
    + depend on portmap 6.0-10ubuntu1.
  - debian/nfs-kernel-server.postinst: don't call "invoke-rc.d nfs-common"
    in the postinst, this is redundant anyway and the nfs-common init script
    is gone now.
* Merge from Debian unstable, remaining changes:
  - debian/nfs-common.{statd,gssd,idmapd,rpc_pipefs}.upstart,
    debian/control, debian/nfs-common.{preinst,postinst,prerm,postrm},
    debian/rules: drop nfs-common init script in favor of upstart jobs, and
    build-depend on debhelper (>= 7.3.15ubuntu3) for correct upstart init
    handling.
  - debian/control: depend on the upstart-using version of portmap,
    6.0-10ubuntu1; and drop the alternative depends on rpcbind, which
    hasn't been converted.
  - debian/nfs-kernel-server.postinst: don't call "invoke-rc.d nfs-common"
    in the postinst, this is redundant anyway and the nfs-common init script
    is gone now.
* Dropped changes, included in Debian:
  - debian/control: add ${misc:Depends} line for nfs-common, for the
    upstart deps
  - debian/nfs-kernel-server.init: updated check for presence of nfsd
    support
* debian/nfs-common.gssd.upstart:
  - fix the OPTIONS= match for the start condition (missing a leading '*')
    so that we actually match on nfs4 mounts
  - drop the 'script' for a straight exec of rpc.gssd; if /usr is a separate
    partition then nfs4 mounts might be attempted in parallel, and upstart
    gets mightily confused when this happens. LP: #545673
* debian/nfs-common.*.upstart: start on new 'mounting' signal instead of
  obsolete 'mount' signal.
* debian/nfs-kernel-server.init: 2.6.32 kernels no longer export the
  same symbols. Switch symbol check to nfsd_serv which has been present
  since 2005. LP: #493145.
* Merge from Debian testing, remaining changes:
  - debian/nfs-common.{statd,gssd,idmapd,rpc_pipefs}.upstart,
    debian/control, debian/nfs-common.{preinst,postinst,prerm,postrm},
    debian/rules: drop nfs-common init script in favor of upstart jobs, and
    build-depend on debhelper (>= 7.3.15ubuntu3) for correct upstart init
    handling.
  - debian/control: depend on the upstart-using version of portmap,
    6.0-10ubuntu1; and drop the alternative depends on rpcbind, which
    hasn't been converted.
  - debian/control: add ${misc:Depends} line for nfs-common, for the
    upstart deps
  - debian/nfs-kernel-server.postinst: don't call "invoke-rc.d nfs-common"
    in the postinst, this is redundant anyway and the nfs-common init script
    is gone now.
* debian/nfs-common.statd.upstart: check for a started portmap in a
  non-racy manner. LP: #484209.
* debian/control: add missing ${misc:Depends} for nfs-common, else we don't
  get the dependency on upstart that we should have. LP: #456281.
* debian/rules: now that the jobs will exit cleanly on their own when
  'start' is called but the job is a no-op, remove the --error-handler
  option to dh_installinit so that we don't accidentally ignore other
  kinds of errors that could point to real problems.
* Drop the gssd upstart job's dependency on "local-filesystems"; at boot
  time this is always implied transitively by the dep on portmap, and using
  a combination of 'or' and 'and' operators in the dependency list seems
  to confuse upstart quite badly, causing kerberized mounts to hang at boot.
  LP: #447654.
* Set upstart jobs to also start on mount attempt, in the event that
  mountall gets to them before the daemons are done starting. Really-fixes
  LP: #431248.
* Call 'stop' in the pre-start scripts for all jobs when we want to prevent
  the job from starting; this lets upstart know that it's a clean stop,
  and avoids boot-time messages about service start failures
* Mounting rpc_pipefs also requires the sunrpc module, so move this
  modprobe to the right upstart job.
* nfs-kernel-server: don't call invoke-rc.d nfs-common in the postinst,
  this is redundant anyway and the nfs-common init script is gone now.
  LP: #441855.
* Configure gssd and idmapd upstart jobs to stop on runlevels 0 and 6; this
  is consistent with previous initscript-based behavior, and spares upstart
  trying to restart the jobs repeatedly when sendsigs runs.
* When autodetecting gssd, handle the case of 'sec=krb5' being embedded in
  the middle of the options list in /etc/fstab. LP: #364861.
* Fix transition idempotency error when stopping old nfs-common init
  script, in case the postinst fails to finish afterwards (e.g., failure to
  restart the daemons).
* Fix statd upstart job to properly honor NEED_STATD=no and not get stuck
  respawning indefinitely.
* Ignore failures to start the daemons on upgrade, since if they aren't
  needed we don't *want* them to start. LP: #441055.
* Fix up the rpc_pipefs job, the wrong version of the file slipped into
  the previous upload.
* Drop nfs-common init script in favor of new upstart jobs. LP: #431248.
* Build-depend on debhelper (>= 7.3.15ubuntu3) for correct upstart init
  handling.
* Depend the upstart-using version of portmap, 6.0-10ubuntu1; and drop the
  alternative depends on rpcbind, which hasn't been converted.

lp:~ubuntu-branches/ubuntu/natty/nfs-utils/natty-201101190615 (Has a merge proposal) 1 Development 2011-01-19 06:15:58 UTC
40. roll back the gssd fixes; there's a r...

Author: Steve Langasek
Revision Date: 2011-01-19 01:42:31 UTC

roll back the gssd fixes; there's a race condition here as well where an NFS
'mounting' hook can be sent before local-filesystems is up, and we can't fix
that without deadlocking boot until mountall is fixed.

16 of 6 results