Releases: tarantool/cartridge
2.12.4 - bumped dependencies
2.12.3 - issues improvements
Added
-
VShard alerts can be displayed in issues list after enabled with env/cli args.
-
New option
TARANTOOL_DISABLE_UNRECOVERABLE_INSTANCES
to automatically disable
instances with stateInitError
orBootError
.
Fixed
- Last instance in vshard-storage replicaset can be expelled now.
Changed
-
Cartridge doesn't use
vshard-ee
instead ofvshard
. -
Cartridge doesn't use
ddl-ee
instead ofddl
. -
uuids in issues replaces with instance names and uris.
2.12.2 - another membership bug fixed
Changed
- Update
membership
dependency to 2.4.5.
Added
-
cartirdge.get_opts()
function to get current Cartridge options. -
More logging cartridge options on start.
2.12.1 - bugfix for ee modules support
Fixed
ddl-manager-ee
is now in the list of implicit roles.
Added
auth_enabled
paramtest-helpers.cluster
to enable/disable auth in tests.
2.12.0 - EE modules support
Added
-
VShard and DDL versions are displayed in the WebUI.
-
cartridge.cfg
paramset_cookie_hash_membership
to set
cluster cookie hash as encryption key in membership instead of
plain cookie.
Changed
-
Cartridge now uses
vshard-ee
instead ofvshard
if available. -
Cartridge now uses
ddl-ee
instead ofddl
if available.
Deprecated
- Usage of plain cluster cookie as a membership encryption key by default.
2.11.0 - introduce disk failure check
Added
-
Disk failure check. If there is something wrong with the disk, the instance
will be disabled automatically and the corresponding issue will be shown
in the WebUI. -
GraphQL API to enable previously disabled instances:
mutation { cluster { enable_servers(uuids: [...]) { } } }
.
Changed
-
Disabling/enabling instances with
disable_servers
/enable_servers
mutations automatically disables/enables VShard storages. -
Update
ddl
dependency to 1.7.1. -
Update
vshard
dependency to 0.1.27.
Fixed
-
Fix false positive warning in migrations UI.
-
Leader autoreturn doesn't try to return leadership to unhealthy leader anymore.
2.10.0 - new maintenance features in API and WebUI
Added
-
Add the state provider status action on the failover controller modal.
-
Add the "Migrations" page.
-
Add "rebalancer_mode" and "rebalancer" options on web UI.
-
Twophase commit timeouts now can be set with env.
-
New GraphQL API
failover_state_provider_status
to ping state provider connection. -
New issue about unhealthy replicasets.
Changed
-
More strict validation for
cartridge.is_healthy
API function. -
Update
membership
dependency to 2.4.4. -
Update
ddl
dependency to 1.7.0.
Fixed
- Switching leadership when disabling leader in replicaset.
2.9.0 - introduce rebalancer management
Fixed
-
Raft failover state transitions.
-
Remote control socket doesn't bind to a localhost when different host is available.
Added
-
New VShard configuration options:
rebalancer
(on server/replicaset level)
andrebalancer_mode
(on VShard config level). -
rebalancer_enabled
field to boxinfo GraphQL API.
Changed
- Config section names with relative paths are prohibited.
2.8.6 - bugfix release
2.8.5 - more maintenance features and IPv6 support
Added
-
election_leader_idle
field to GraphQL API. -
new issue when
box.info.election.leader_idle
is too high. -
Lua API
get_uris
to get all instances uris. -
Filter param to Lua API
get_servers
to filter instances. -
new issue about vshard storages marked as
ALL_RW
. -
cartridge.cfg
optiondisable_raft_on_small_clusters
to disable Raft
failover on clusters with less than 3 instances (default:true
). -
argparse
now logs if some sections in config files were ignored (#2169). -
IPv6 support (#2166).
Fixed
- An error with
cartridge.utils.version_is_at_least
parsing.