summaryrefslogtreecommitdiffstats
path: root/xlators/mgmt/glusterd/src/glusterd-mgmt.c
Commit message (Collapse)AuthorAgeFilesLines
* glusterd/snapshot: Recording the snapshots missed in each brick.Avra Sengupta2014-04-021-9/+14
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Persisting missing snapshot info on disk as well as in memory in the following format: -------------NODE-UUID--------------:---------SNAP-UUID--------------=BRICKNUM:-------BRICKPATH--------:OPERATION:STATUS 927cb5fe-63da-48f5-82f6-e6a09ddc81c4:a17b4fe42c5a45f7a916438643edaa13= 3 :/brick/brick-dirs/brick3: 1 : 1 927cb5fe-63da-48f5-82f6-e6a09ddc81c4:a17b4fe42c5a45f7a916438643edaa13= 3 :/brick/brick-dirs/brick3: 3 : 1 927cb5fe-63da-48f5-82f6-e6a09ddc81c4:83a3cc05453b46b2a7eda4c9a9208638= 3 :/brick/brick-dirs/brick3: 1 : 1 This data will be stored on disk at /var/lib/glusterd/snaps/missed_snaps_list In memory we maintain the data as a list of glusterd_missed_snap_info in conf, the key for this list are the first two fields, i.e NODE-UUID:SNAP-UUID. For every NODE-UUID:SNAP-UUID, there can be multiple operations missed on multiple bricks. So we maintain a list of glusterd_snap_op_t for evert node of glusterd_missed_snap_info This list is maintained or updated during snapshot create, delete, and restore operations which are the only operations that if missed, are recorded in this list. During snapshot create, if a node is down, or a brick is down, we don't receive their mount point infos. snap_status of such bricks is marked as -1, and their brick details are added to this list. During snapshot delete, we check from originator node, if any other nodes, holding bricks of the said snap are down. Those are also added to the list. Also if the node is up, but the snapshot was pending for a snap brick, and its snap_status is -1, we add that to the list too. When a subsequent delete entry is processed for an already existing create entry, we just mark the create entries status as done (2), and don't add the delete entry to the list. During snapshot restore, we check from originator node, if any other nodes, holding bricks of the said snap are down. Those are also added to the list. Also if the node is up, but the snapshot was pending for a snap brick, and its snap_status is -1, we add that to the list too. Change-Id: I22578d14f81a54e13f6832966b70cd4cfdfd5b44 Signed-off-by: Avra Sengupta <asengupt@redhat.com> Reviewed-on: http://review.gluster.org/7208 Reviewed-by: Vijaikumar Mallikarjuna <vmallika@redhat.com> Reviewed-by: Rajesh Joseph <rjoseph@redhat.com> Tested-by: Rajesh Joseph <rjoseph@redhat.com>
* Crash Fixes introduced by review commentsAvra Sengupta2014-03-281-4/+0
| | | | | | | | | | Change-Id: I22f9828e39494a8e1ad37c9a2ff78866d117e6af Signed-off-by: Avra Sengupta <asengupt@redhat.com> Reviewed-on: http://review.gluster.org/7365 Reviewed-by: Sachin Pandit <spandit@redhat.com> Reviewed-by: Vijaikumar Mallikarjuna <vmallika@redhat.com> Reviewed-by: Rajesh Joseph <rjoseph@redhat.com> Tested-by: Rajesh Joseph <rjoseph@redhat.com>
* Review CommentsAvra Sengupta2014-03-271-154/+305
| | | | | | | | Change-Id: Ifce90b0b617bc0b43a9af0bd692a7290820ac62c Signed-off-by: Avra Sengupta <asengupt@redhat.com> Reviewed-on: http://review.gluster.org/7358 Reviewed-by: Rajesh Joseph <rjoseph@redhat.com> Tested-by: Rajesh Joseph <rjoseph@redhat.com>
* glusterd/mgmt_v3_locks: Send CLI response only after doing unlockVijaikumar M2014-03-251-35/+36
| | | | | | | | | Change-Id: Ibeaa1377b904430fabea5dec2d38b87c1bd52161 BUG: 1077196 Signed-off-by: Vijaikumar M <vmallika@redhat.com> Reviewed-on: http://review.gluster.org/7288 Reviewed-by: Rajesh Joseph <rjoseph@redhat.com> Tested-by: Rajesh Joseph <rjoseph@redhat.com>
* mgmt/glusterd: do cleanup of snapshots in post-validate phase if half bakedRaghavendra Bhat2014-03-241-34/+132
| | | | | | | | | | objects are there Change-Id: I372cac98ad054cdc1a6fbc7f6c77c25981063b2f Signed-off-by: Raghavendra Bhat <raghavendra@redhat.com> Reviewed-on: http://review.gluster.org/7237 Reviewed-by: Rajesh Joseph <rjoseph@redhat.com> Tested-by: Rajesh Joseph <rjoseph@redhat.com>
* glusterd/mgmt_v3 locks: Using snaps as an entity for mgmt_v3 locks.Avra Sengupta2014-03-071-52/+26
| | | | | | | | | | | | Created framework to process single volume lock, single snap lock, multiple volume locks, multiple snap locks, and multiple snap and volume locks requests. Change-Id: I3f87bdec0bf142d6a1fd3f9d9f5012cc44eaf4bd Signed-off-by: Avra Sengupta <asengupt@redhat.com> Reviewed-on: http://review.gluster.org/7189 Reviewed-by: Rajesh Joseph <rjoseph@redhat.com> Tested-by: Rajesh Joseph <rjoseph@redhat.com>
* glusterd/mgmt_v3 locks: Handling different entities with mgmt_v3_locks.Avra Sengupta2014-03-071-3/+3
| | | | | | | | | | | | | | | | | Within the same namespace allowing mgmt_v3 lock engine to handle entities of valid types. Postfixing the element's name with the entity type allows us to perform locking operations on different elements of different entitites at the same time. For example: A volume named test, will acquire a lock on the name test_vol, while a snap named test will acquire lock on the name test_snap, and thus be allowed to co-exist in the same namespace. Change-Id: I611d1a59b707e497d35a37bdca5a66e29a1b6f06 Signed-off-by: Avra Sengupta <asengupt@redhat.com> Reviewed-on: http://review.gluster.org/7188 Reviewed-by: Rajesh Joseph <rjoseph@redhat.com> Tested-by: Rajesh Joseph <rjoseph@redhat.com>
* glusterd/mgmt_v3 locks: Generalization of the volume wide locks.Avra Sengupta2014-03-071-64/+62
| | | | | | | | | | Renaming volume locks as mgmt_v3 locks Change-Id: I2a324e2b8e1772d7b165fe96ce8ba5b902c2ed9a Signed-off-by: Avra Sengupta <asengupt@redhat.com> Reviewed-on: http://review.gluster.org/7187 Reviewed-by: Rajesh Joseph <rjoseph@redhat.com> Tested-by: Rajesh Joseph <rjoseph@redhat.com>
* glusterd: Pass generated snap-name/cg-name to cli.Avra Sengupta2014-01-091-9/+26
| | | | | | | | Also block operation on response aggregation errors. Change-Id: I641ddc98d986fd44627d790ab84488a6b849f807 BUG: 1043792 Signed-off-by: Avra Sengupta <asengupt@redhat.com>
* glusterd/mgmtv3: Fetching snap mount path per brick.Avra Sengupta2013-12-031-4/+63
| | | | | | | | | | | | | | | | | Fetching snap mount path per brick during pre-validate and aggregating it before commit to generate all brick volfiles and all client volfiles on all nodes. Added changes to propagate the same cg-id by generating it in pre-validate and sending the same to all the nodes during commit Also fixed volume-id mismatches caused in snap-volumes leading to failures of volume status Change-Id: I0b8ed81ac93077cd50313e6eeed00f71932cf2ac Signed-off-by: Avra Sengupta <asengupt@redhat.com> Signed-off-by: Rajesh Joseph <rjoseph@redhat.com>
* mgmt/glusterd: handle issues present in snapshot create and snap managementRaghavendra Bhat2013-12-021-28/+6
| | | | | Change-Id: I94b5f6e00be7d1ff0c454e291c779dae7b423748 Signed-off-by: Raghavendra Bhat <raghavendra@redhat.com>
* glusterd/snapshot: Added trace statements and handled snapshot create commit ↵Avra Sengupta2013-11-201-10/+14
| | | | | | | | | | | failure. Also handles empty string(not NULL) in gd_syncop_mgmt_brick_op() and adds "Snapshot" in operation list used for printing op during logging. Change-Id: Icac9dce6bf1c087ab2aace9953e2af3a0fb81be6 Signed-off-by: Avra Sengupta <asengupt@redhat.com>
* glusterd/mgmt: Adding synctasked flag to initiate multiple volume locks.Avra Sengupta2013-11-201-0/+8
| | | | | | | | Setting 'is_synctasked' as gf_true for glusterd_mgmt_v3_initiate_snap_phases() to initiate multiple volume locks. Change-Id: Ib225fbac42b5e9cf9f3af0083d33ec49a1b87964 Signed-off-by: Avra Sengupta <asengupt@redhat.com>
* glusterd/Jarvis: Added aggr rsp dict in mgmt frameworkAvra Sengupta2013-11-151-3/+45
| | | | | | | Also fixes snapshot config output Change-Id: Ia50d94492009cf73dbb99ba20117b9fa4c41048a Signed-off-by: Avra Sengupta <asengupt@redhat.com>
* snapshot: Snapshot restoreRajesh Joseph2013-11-151-0/+25
| | | | | | | | | | | | GL-31: Ability to restore snapshot Implemented snapshot restore for thin logical volume. As of now snapshot restore for CG is not tested. Testing for snapshot restore of a volume is done by changing the snapshot create process to create a thick snapshot. This is done because --merge option to restore thin volume is not working in the latest kernel. Change-Id: Ia3ded7e6c4da5957a74e269a25ba3200e6fb2d8b Signed-off-by: Rajesh Joseph <rjoseph@redhat.com>
* mgmt/glusterd: snapshot config changesshishir gowda2013-11-151-8/+0
| | | | | | | | Also refactored code in glusterd for create command Additionally, removed brick-op func from mgmt_iniate_all_phases Change-Id: Iddcc332009c5716adee7f2b04c93b352fb983446 Signed-off-by: shishir gowda <sgowda@redhat.com>
* mgmt/snapshot: brick op for starting/stopping barriershishir gowda2013-11-151-1/+200
| | | | | Change-Id: Iafbd0ec95de0c41455fb79953fb4bb07721334a5 Signed-off-by: shishir gowda <sgowda@redhat.com>
* mgmt/glusterd: snapshot create commandRaghavendra Bhat2013-11-151-7/+33
| | | | | | | | | | | | | | | | | This is still a work in progress. As of now, these things are done: * Take the snapshot of the backend brick * Create the new volume for the snapshot * Create the brick and the client volfiles * Store the snapshot related info in /var/lib/glusterd * Create the snap object representing the snapshot TODO: Start the brick processes for the snapshot Change-Id: I26fbb0f8e5cf004d4c1dbca51819bab1cd1bac15 Signed-off-by: Raghavendra Bhat <raghavendra@redhat.com>
* glusterd/locks: Adding multiple volume locks supportsAvra Sengupta2013-11-151-76/+93
| | | | | | | Also linking snap create command to mgmt_v3 Change-Id: If2ed29be072e10d0b0bd271d53e48eeaa6501ed7 Signed-off-by: Avra Sengupta <asengupt@redhat.com>
* glusterd: Defining interfaces to make use of mgmt_v3 rpcsAvra Sengupta2013-11-151-0/+1236
| | | | | | | | | Defining separate interfaces for every phase to make use of the rpcs and providing set of integrated interfaces for commands to consume Change-Id: I6d464326c5a8b5875a7c2539c9df072b23fe61a9 Signed-off-by: Avra Sengupta <asengupt@redhat.com>
* glusterd/mgmt_v3: Initating complete synctask on glusterdAvra Sengupta2013-11-151-0/+67
glusterd mgmt_v3 is nothing but a complete synctask approach for glusterd to function. The commands making use of this won't be using the op-state machine to inject events and will be using the synctask framework to perform operations across all nodes in the cluster. This patch defines the program and the handlers used. Change-Id: Ibff2c62b0187c40cdea7254c85786297bba60372 Signed-off-by: Avra Sengupta <asengupt@redhat.com>
generated by cgit (git 2.34.1) at 2024-04-23 23:17:46 +0000