* Consolidated _servers and _serverAdvertisedEndpoints, added rebootId, prepared change notifications
* Cleanup
* Added a RebootId type
* Began implementing RebootTracker (still WIP)
* Moved RebootId operators into the class
* Removed RebootId operator<< again
* Added tests, added CallbackGuard, removed/commented old RebootTracker code
* Fix: do not try to call unset callbacks
* Split one test, added another
* Added more tests
* Renamed tests, added more tests
* Fixed missing variable declarations
* Let MockServer appear to be started
* Reorded test, fixed naming
* Implemented callMeOnChange()
* Re-implemented RebootTracker (not yet working)
* Resolved a TODO, updated a test, added comments
* Call old callbacks immediately
* Fixed tests
* Use EXPECT_* instead of ASSERT_*
* Suppress a log message
* Resolved TODOs
* Reverted changes on reading ServersRegistered
* Update RebootTracker
* Introduce `rebootId` into ServerState for Cluster
* A server *boots* if it is started on a previously non-existing data
directory and hence does not have a UUID yet.
* A server *reboots* if it is started on a pre-existing data directory
We keep the rebootId in the cluster's agency under
Current/ServersKnown/$uuid/rebootId.
When rebooting (and subsequently re-joining a cluster), the server increments
its rebootId in Phase 2 of registration. This way it can be detected within the
cluster whether a server was restarted.
This information will later be used to handle cases where server restarts can
lead to problems, for example with transactions or in-progress queries.
* Move rebootId into Current/ServersKnown/
* Fixed typo
* Fixed log ids
* Add deletion of ServersKnown/UUID from agency
* Add deletion of Current/ServersKnown/UUID to removeServer
* Clean up readRebootIdFromAgency and add retry loop around it
* Bugfix
* Added nolint comments
* Fixed initialization order
* Fixed ClusterInfo-test
* Added log messages
* Revert "Fixed ClusterInfo-test"
This reverts commit d983596979.
* Disabled assertion for google tests
* Ignore windows compile warning
* Always call loadServers in loadCurrent
* Fix really subtle bug when not returning a value
* Introduce `rebootId` into ServerState for Cluster
* A server *boots* if it is started on a previously non-existing data
directory and hence does not have a UUID yet.
* A server *reboots* if it is started on a pre-existing data directory
We keep the rebootId in the cluster's agency under
Current/ServersKnown/$uuid/rebootId.
When rebooting (and subsequently re-joining a cluster), the server increments
its rebootId in Phase 2 of registration. This way it can be detected within the
cluster whether a server was restarted.
This information will later be used to handle cases where server restarts can
lead to problems, for example with transactions or in-progress queries.
* Move rebootId into Current/ServersKnown/
* Add deletion of ServersKnown/UUID from agency
* Add deletion of Current/ServersKnown/UUID to removeServer
* Clean up readRebootIdFromAgency and add retry loop around it
* Fixed compile error due to forbidden implicit cast
* Fixed compile error on windows
* Fixed compile error due to devel merge
* Removed dead comment
* Removed TODO note
* Extended comment
* Removed TODO note
* Fixed using an invalidated iterator
* Copy string only if necessary
* Fixed compile error
* port of feature-3.4/mv-gzip-export to devel branch
* add explicit namespaces so gcc 6.3.0 would successfully compile
* add conditional cleanup of ENCRYPTION file from another branch
* use _lseek() for Windows build to avoid deprecated warning.
* change from ifdef for lseek variants to TRI_LSEEK.
* force Windows lseek to return Linux expected type.
* Added startup error for bad temporary directory setting.
If the temporary directory (--temp.path) setting is identical to the database
directory (`--database.directory`) this can eventually lead to data loss, as
temporary files may be created inside the temporary directory, causing overwrites of
existing database files/directories with the same names.
Additionally the temporary directory may be cleaned at some point, and this would lead
to an unintended cleanup of the database files/directories as well.
Now, if the database directory and temporary directory are set to the same path, there
will be a startup warning about potential data loss (though in ArangoDB 3.4 allowing to
continue the startup - in 3.5 and higher we will abort the startup).
* Improve collection time of IResearchQueryOptimizationTest
* Added a minReplicationFactor field in Collections. It is not possible to modify it yet and noone cares for it
* Added some assertion son minReplicationFactor
* Transaction API will now reject writes as soon as minimal replication factor is NOT fulfilled
* added minReplicationFactor to the user interface, preparation for the collection api changes
* added minReplicationFactor to VocBaseCollection, RestReplicationHandler, RestCollectionHandler, ClusterMethods, ClusterInfo and ClusterCollectionCreationInfo
* added minReplicationFactor usage to tests
* TODO TEMOPORARY COMMIT FOR TESTING PLEASE REVERT ME
* minReplicationFactor now able to change via collection properties route
* fixed wrongly assert
* added minReplicationFactor to the graph management ui
* added minReplicationFactor to the gharial api
* Fixed off-by-one error in minReplicationFactor. We actually enforced one more.
* adjusted description of minReplicationFactor
* FollowerInfo Refactoring
* added gharial api graph creation tests with minimal replication factor
* proper cleanup of shell collection tests, removed lots of duplicate code, preparation for some new tests
* added collection create tests using invalid/valid names, replicationFactor and minReplicationFactor
* Debug logging
* MORE Debug logging
* Included replication fast lane
* Use correct minreplicationfactor
* modified debug logging
* Fixed compileissues
* MORE Debug logging
* MORE Debug logging
* MORE Debug logging
* MORE Debug logging
* MORE Debug logging
* MORE Debug logging
* MORE Debug logging
* Revert "MORE Debug logging"
This reverts commit dab5af28c0.
* Revert "MORE Debug logging"
This reverts commit 6134b664bd.
* Revert "MORE Debug logging"
This reverts commit 80160bdf3b.
* Revert "MORE Debug logging"
This reverts commit 06aabcdfe1.
* Removed debug output
* Added replication fast lane. Also refactored the commands as i cannot take it any more...
* Put some requests of RocksDBReplication onto CATCHUP Lane.
* Put some requests of MMFilesReplication onto CATCHUP Lane.
* Adjusted Fast and MED lane usage in Supervised scheduler
* Added changelog entry
* Added new features entry
* A new leader will now keep old followers in case of failover
* Update arangod/Cluster/ClusterCollectionCreationInfo.cpp
Co-Authored-By: Tobias Gödderz <tobias@arangodb.com>
* Fixed JSLINT
* Unified lane handling of replication handlers
* Sorry forgotten in last commit
* replaced strings with static strings
* more use of static strings
* optimized min repl description in the ui
* decr initial loop variable
* clean up of the createWithId test
* more use of static strings
* Update js/apps/system/_admin/aardvark/APP/frontend/js/views/collectionsView.js
Co-Authored-By: Tobias Gödderz <tobias@arangodb.com>
* Added some comments on condition, renamed variable as suggested in review
* Added check for min replicationFactor to be non-zero
* Added assertion
* Added function to modify min and max replication factor in one go
* added missing semicolon
* rm log devel
* Added a second information to follower info that can keep track of followers that have been in sync before a failover has taken place
* Maintenance reports previous version now to follower info. instead of lying by itself. The Follower Info now gets a failover save mode to report insync followers
* check replFactor against nr dbservers
* Add lie reporting in CURRENT
* Reverted most of my recent commits about Failover situation. The intended plan simply does not work out
* move replication checks from logical collection to rest collection handler
* added more replication tests
* Include assert only if we are not in gtest
* jslint
* set min repl factor to zero if satellite collection
* check replication attributes in v8 collection
* Initial commit, old plan, does not yet work
* fixed ires tests
* Included FailoverCandidates key. Not fully implemented
* fixed wrong assert
* unified in sync follower reporting
* fixed compiler errors
* Cleanup locking, and fixed potential deadlocks
* Comments about locking order in FollowerInfo.
* properly check uint
* Keep old leader as potential failover candidate
* Transaction methods now use followerInfo to check if the leader can write, this might have the sideeffect that 'failoverCandidates' are updated
* Let agency check failoverCandidates if possible
* Initialize member variables
* Use unified follower reporting in DBServerAgencySync
* Removed obsolete variable, collecting it somewhere else
* repl factor attr check
* Reimplemented previous followers, second attempt now. PhaseOne and PhaseTwo can now synchronize on current.
* Fixed assertion, forgot an off-by-one
* adjusted test to be more preciese now
* Fixed failove candidates list
* Disable write on dropping too many followers
* Allow to run updateFailoerCandidates multiple times with same leader.
* Final fixes, resilience tests now green, crossing fingers for jenkins
* Fixed race on atomics comparison
* Fixed invalid number type
* added nullptr handling
* added nullptr handling
* Removed invalid assert
* Make takeover of leadership an atomic operation
* Update tests/js/common/shell/shell-cluster-collection.js
Co-Authored-By: Tobias Gödderz <tobias@arangodb.com>
* Review fixes
* Fixed creation code to use takeoverLeadership
* Update arangod/Cluster/FollowerInfo.h
Co-Authored-By: Tobias Gödderz <tobias@arangodb.com>
* Applied review fixes
* There is no timeout
* Moved AQL + Pregel to INTERNAL_AQL lane, which is medium priority, to avoid deadlocks with Sync replication
* More review fixes
* Use difference if you want to compare two vectors...
* Use std::string ...
* Now check if we are in recovery mode
* Added documentation for minReplicationFactor
* Added readme update as well in documenation
* Added a minReplicationFactor field in Collections. It is not possible to modify it yet and noone cares for it
* Added some assertion son minReplicationFactor
* Transaction API will now reject writes as soon as minimal replication factor is NOT fulfilled
* added minReplicationFactor to the user interface, preparation for the collection api changes
* added minReplicationFactor to VocBaseCollection, RestReplicationHandler, RestCollectionHandler, ClusterMethods, ClusterInfo and ClusterCollectionCreationInfo
* added minReplicationFactor usage to tests
* TODO TEMOPORARY COMMIT FOR TESTING PLEASE REVERT ME
* minReplicationFactor now able to change via collection properties route
* fixed wrongly assert
* added minReplicationFactor to the graph management ui
* added minReplicationFactor to the gharial api
* Fixed off-by-one error in minReplicationFactor. We actually enforced one more.
* adjusted description of minReplicationFactor
* FollowerInfo Refactoring
* added gharial api graph creation tests with minimal replication factor
* proper cleanup of shell collection tests, removed lots of duplicate code, preparation for some new tests
* added collection create tests using invalid/valid names, replicationFactor and minReplicationFactor
* Debug logging
* MORE Debug logging
* Included replication fast lane
* Use correct minreplicationfactor
* modified debug logging
* Fixed compileissues
* MORE Debug logging
* MORE Debug logging
* MORE Debug logging
* MORE Debug logging
* MORE Debug logging
* MORE Debug logging
* MORE Debug logging
* Revert "MORE Debug logging"
This reverts commit dab5af28c0.
* Revert "MORE Debug logging"
This reverts commit 6134b664bd.
* Revert "MORE Debug logging"
This reverts commit 80160bdf3b.
* Revert "MORE Debug logging"
This reverts commit 06aabcdfe1.
* Removed debug output
* Added replication fast lane. Also refactored the commands as i cannot take it any more...
* Put some requests of RocksDBReplication onto CATCHUP Lane.
* Put some requests of MMFilesReplication onto CATCHUP Lane.
* Adjusted Fast and MED lane usage in Supervised scheduler
* Added changelog entry
* Added new features entry
* A new leader will now keep old followers in case of failover
* Update arangod/Cluster/ClusterCollectionCreationInfo.cpp
Co-Authored-By: Tobias Gödderz <tobias@arangodb.com>
* Fixed JSLINT
* Unified lane handling of replication handlers
* Sorry forgotten in last commit
* replaced strings with static strings
* more use of static strings
* optimized min repl description in the ui
* decr initial loop variable
* clean up of the createWithId test
* more use of static strings
* Update js/apps/system/_admin/aardvark/APP/frontend/js/views/collectionsView.js
Co-Authored-By: Tobias Gödderz <tobias@arangodb.com>
* Added some comments on condition, renamed variable as suggested in review
* Added check for min replicationFactor to be non-zero
* Added assertion
* Added function to modify min and max replication factor in one go
* added missing semicolon
* rm log devel
* Added a second information to follower info that can keep track of followers that have been in sync before a failover has taken place
* Maintenance reports previous version now to follower info. instead of lying by itself. The Follower Info now gets a failover save mode to report insync followers
* check replFactor against nr dbservers
* Add lie reporting in CURRENT
* Reverted most of my recent commits about Failover situation. The intended plan simply does not work out
* move replication checks from logical collection to rest collection handler
* added more replication tests
* Include assert only if we are not in gtest
* jslint
* set min repl factor to zero if satellite collection
* check replication attributes in v8 collection
* fixed ires tests
* fixed wrong assert
* properly check uint
* repl factor attr check
* adjusted test to be more preciese now
* Fixed race on atomics comparison
* Fixed invalid number type
* Update tests/js/common/shell/shell-cluster-collection.js
Co-Authored-By: Tobias Gödderz <tobias@arangodb.com>
* Review fixes
* More review fixes
* make rspec timeouteable (on windows for now)
* add timeout state, fix windows implementation
* abort testrun if a timeout in rspec occured
* lint
* fix scope of procdump stopper
* fix include scope
* fix invokation
* use proper variable to wait for
* Bug fix 3.4/collection babies (#9033)
* Prepare API to create multiple collections in a single request to ClusterMethods to improve speedup
* Added counter on how many collections are successfully created
* Allow multi collection creation one level higher
* CollectionMethods now allow batch createion of Collections
* Improved array size assertions
* Now a graph is createad within a single roundtrip in the agency.
* Added new header files
* Insert collections in the AGENCY with TTL and a isBuilding flag, collections with this flag should not be visisible in the coordinator
* Added forgotten C++ file
* Fixed a rare race condition, and the failing IResearch Tests
* readded callback on DONE, otherwise lists are out of sync
* Fixed assertions to let mocked tests pass...
* Fixed community cluster
* Started fixing IResearch analyzer test, catch-tests are failing ;(
* Solved missed merge-conflict
* Added helper functions in AnalyzerFeature-test
* Refactoring AnalyzerTest Section-Auth
* Refactoring AnalyzerTest Section-Emplace-Duplicates
* Refactoring AnalyzerTest Section-Emplace-Error-Cases. Recovery-Test is now red, it seemed to be green because of invalid test case before.
* Refactoring AnalyzerTest, split GET test into multiple parts, still left 'cluster simulation'.
* Attempt to extract Coordinator / DBServer tests a little bit. This commit starts to break all Coordinator tests. However i am convinced that earlier version did NOT test a cluster situation at all, but some hybrid of SingleServer with full local storage that got told to be a Coordinator from now on, but without any Coordinator setup...
* Temporarly disabled some tests in AnalyzerFeature, as discussed with @gnusi.
* Fixed include guard.
* Temporarily deactivated failing tests
* You shall save your files before you commit...
* Fixed test asserting on plan version, which is now higher than before
* Wait for replication before inserting documents
Also, increased some timeouts and fixed a log message
* Fixed some log levels and a log message
* Removed repair-distribute-shards-like-spec from greylisted tests
* Unbreak aardvark when --foxx.api false
The easiest way to disable the aardvark foxx routes is to just skip the
import and mount altogether. The previous approach caused the service to
throw an error because the module doesn't actually export anything.
There's no need to prevent the export if we just don't import the module
to begin with.
* Add an error code for service api disabled
* Throw an actual error when accessing Foxx API routes
* Throw an error when using legacy Foxx API
There's no easy way to use a catch-all here so disable them one by one. No new routes will be added to this API and it'll go away in 4.0 so there's no risk of forgetting to add the throw-check in new routes.
* Render subnav without delay
The subnav doesn't need the perm check results but we need the subnav to exist so we can populate it in other parts of the app. This fixes the subnav not showing up when reloading certain views (e.g. install service via upload).
* Hide ONLY the Foxx store when it is disabled
We still want to allow installing and upgrading/replacing services. We just don't want to show the store.
* Hide Foxx services in frontend if Foxx API is disabled
* Build aardvark
* fixed typo in disable check, fixed tests
* Unbreak aardvark when --foxx.api false
The easiest way to disable the aardvark foxx routes is to just skip the
import and mount altogether. The previous approach caused the service to
throw an error because the module doesn't actually export anything.
There's no need to prevent the export if we just don't import the module
to begin with.
* Add an error code for service api disabled
* Throw an actual error when accessing Foxx API routes
* Throw an error when using legacy Foxx API
There's no easy way to use a catch-all here so disable them one by one. No new routes will be added to this API and it'll go away in 4.0 so there's no risk of forgetting to add the throw-check in new routes.
* Hide Foxx services in frontend if Foxx API is disabled
* Render subnav without delay
The subnav doesn't need the perm check results but we need the subnav to exist so we can populate it in other parts of the app. This fixes the subnav not showing up when reloading certain views (e.g. install service via upload).
* Hide ONLY the Foxx store when it is disabled
We still want to allow installing and upgrading/replacing services. We just don't want to show the store.
* Build aardvark
* issue 526.6: implement REST and V8 handlers for the iresearch analyzer feature
* address typo
* remove excess comments
* temporarily comment out tests failing on MacOS
* temporarily comment out more MacOS-only test failures
Simon
Jan
jsteemann
Dan Larkin-York
Wilfried Goesgens
Tobias Gödderz
Frank Celler
Lars Maier
Matthew Von-Maszewski
Michael Hackstein
Andrey Abramov
Jan Christoph Uhde
Markus Pfeiffer
Alan Plum
Frank Celler
Kaveh Vahedipour
Vasiliy
KVS85
Simran
Manuel Pöter