From 689a2374009b41d0790265cb277bb924ffa5cff5 Mon Sep 17 00:00:00 2001 From: Niels de Vos Date: Wed, 27 Apr 2016 10:11:51 +0200 Subject: Move Gluster/NFS disabled by default to 'done/3.8' Change-Id: I192df0b5e42f8f442bc674f46b8f4405e58d6ef9 Signed-off-by: Niels de Vos Reviewed-on: http://review.gluster.org/14087 Reviewed-by: Kaleb KEITHLEY --- accepted/gluster-nfs-off.md | 117 --------------------------------- done/GlusterFS 3.8/gluster-nfs-off.md | 119 ++++++++++++++++++++++++++++++++++ 2 files changed, 119 insertions(+), 117 deletions(-) delete mode 100644 accepted/gluster-nfs-off.md create mode 100644 done/GlusterFS 3.8/gluster-nfs-off.md diff --git a/accepted/gluster-nfs-off.md b/accepted/gluster-nfs-off.md deleted file mode 100644 index 007d9fd..0000000 --- a/accepted/gluster-nfs-off.md +++ /dev/null @@ -1,117 +0,0 @@ -Feature -------- - -Gluster NFS Defaults to Off. - -Summary -------- - -Currently, when a volume is created and started, the default setting for nfs.disable -is off/false; when the volume is started, a glusterfs nfs server process is started. - -The Gluster NFS server is NFSv3 only, and is not being actively developed. Meanwhile -NFS-Ganesha supprts NFSv3, NFSv4, NFSv4.1, and pNFS; and NFSv4.2 is under development. - -Initial support for NFS-Ganesha first appeared in GlusterFS 3.7 with the introduction -of CLI support for using ganesha.nfsd instead of the glusterfs nfs server, however -the glusterfs nfs server remained enabled by default. - -The next step toward eventual deprecation of the glusterfs nfs server is to change the -default setting for nfs.disable to on/true. Users who wish to continue to use the -glusterfs nfs server will have to explicitly set nfs.disable to off/false after they -create the volume. Existing volumes that had nfs.disable set to off/false will not be -altered, those volumes will continue to start the glusterfs nfs server. - -Eventually the glusterfs nfs server will be fully deprecated in favor of NFS-Ganesha. - -Owners ------- - -Kaleb KEITHLEY -Niels De Vos - -Current status --------------- - -Under development - -Related Feature Requests and Bugs ---------------------------------- - -None. - -Detailed Description --------------------- - -None - -Benefit to GlusterFS --------------------- - -Need to maintain the NFS server xlator is eliminated. - - -Scope ------ - -#### Nature of proposed change - -Change default option value in xlators/mgmt/glusterd/... - -#### Implications on manageability - -None - -#### Implications on presentation layer - -None - -#### Implications on persistence layer - -None - -#### Implications on 'GlusterFS' backend - -None - -#### Modification to GlusterFS metadata - -None - -#### Implications on 'glusterd' - -None - -How To Test ------------ - -1. Start glusterd -2. Create volume e.g. `gluster volume create $vol $host:$path [force]` -3. Start volume, e.g. `gluster volume start $vol` -4. check ps to see that glusterfs nfs process is not started - -User Experience ---------------- - -None - -Dependencies ------------- - -None - -Documentation -------------- - -Change any reference to default value of nfs.disable - -Status ------- - -One line change to .../xlators/mgmt/glusterd/src/glusterd-nfs-svc.c has been -posted for review at http://review.gluster.org/13738 - - -Comments and Discussion ------------------------ - diff --git a/done/GlusterFS 3.8/gluster-nfs-off.md b/done/GlusterFS 3.8/gluster-nfs-off.md new file mode 100644 index 0000000..6d37976 --- /dev/null +++ b/done/GlusterFS 3.8/gluster-nfs-off.md @@ -0,0 +1,119 @@ +Feature +------- + +Gluster NFS Defaults to Off. + +Summary +------- + +Currently, when a volume is created and started, the default setting for nfs.disable +is off/false; when the volume is started, a glusterfs nfs server process is started. + +The Gluster NFS server is NFSv3 only, and is not being actively developed. Meanwhile +NFS-Ganesha supprts NFSv3, NFSv4, NFSv4.1, and pNFS; and NFSv4.2 is under development. + +Initial support for NFS-Ganesha first appeared in GlusterFS 3.7 with the introduction +of CLI support for using ganesha.nfsd instead of the glusterfs nfs server, however +the glusterfs nfs server remained enabled by default. + +The next step toward eventual deprecation of the glusterfs nfs server is to change the +default setting for nfs.disable to on/true. Users who wish to continue to use the +glusterfs nfs server will have to explicitly set nfs.disable to off/false after they +create the volume. Existing volumes that had nfs.disable set to off/false will not be +altered, those volumes will continue to start the glusterfs nfs server. + +Eventually the glusterfs nfs server will be fully deprecated in favor of NFS-Ganesha. + +Owners +------ + +* Kaleb KEITHLEY +* Niels de Vos + +Current status +-------------- + +Complete. + +Related Feature Requests and Bugs +--------------------------------- + +[Bug 1092414](https://bugzilla.redhat.com/1092414) + +Detailed Description +-------------------- + +Gluster/NFS should not be running on a clean installation of GlusterFS 3.8. +Only when volumes have `nfs.disable` set to `false` the process should get +started. + +Benefit to GlusterFS +-------------------- + +Need to maintain the NFS server xlator is reduces. It is planned to separate +Gluster/NFS from the standard GlusterFS installation and supply it as an +optional component. Marking the Gluster/NFS service disabled by default is a +step to inform users about the upcoming change, without forcing them to move to +NFS-Ganesha immediately. + + +Scope +----- + +#### Nature of proposed change + +Change default option value in `xlators/mgmt/glusterd/...` and all test-cases +that expect Gluster/NFS to be running by default. + +#### Implications on manageability + +None + +#### Implications on presentation layer + +None + +#### Implications on persistence layer + +None + +#### Implications on 'GlusterFS' backend + +None + +#### Modification to GlusterFS metadata + +None + +#### Implications on 'glusterd' + +None + +How To Test +----------- + +1. Start glusterd +2. Create volume e.g. `gluster volume create $vol $host:$path [force]` +3. Start volume, e.g. `gluster volume start $vol` +4. check ps to see that glusterfs nfs process is not started + +User Experience +--------------- + +Users are advised to deploy NFS-Ganesha instead of Gluster/NFS. However +Gluster/NFS is still available and can easily be enabled by toggling the +`nfs.disable` volume option to `false`. + +Dependencies +------------ + +None + +Documentation +------------- + +Change any reference to default value of `nfs.disable`. + +Comments and Discussion +----------------------- + -- cgit