What's new

Server

• [Backup] do not cancel running backup even if task was moved to another node
• [Cluster] fixed issue with Cluster might accepting commands from higher version than the current one when doing mixed version clusters
• [Counters] remove leading whitespace from Counter Values Change Vector
• [ETL] properly validate if number of transform scripts is greater than 0
• [ETL] properly handling casing mismatch between table name and table definition
• [ETL] state should be tracked per database ID instead of node tag to avoid progress miscalculations when node is being removed and re-added
• [Import] handle compare exchange values incrementally when doing an import from v4.x
• [Indexes] validate if index was created using our database and not copied from a different one
• [Indexes] properly handle cancellation when doing flush and sync to avoid generating index error
• [Monitoring] exposed MemInfo (linux) to SNMP
• [Monitoring] exposed number of faulted databases to SNMP
• [Replication] do not stop replication when replicating a document with a changed collection and there is a conflict already
• [Replication] properly handling Tombstones for legacy 3.x replication
• [Replication] properly setting HasAttachment, HasCounters flag in reverted revision documents
• [Transactions] properly handling Change Vector when TxMerger executes transactions one by one to avoid returning invalid Change Vector
• [Voron] fixed invalid handling of journals created based on recyclable ones during the storage recovery on startup
• [Voron] fixed 'Collection was modified; enumeration operation may not execute.' when generating a storage report
• [.NET Core] updated to .NET Core 3.1.21

Client API

• [Session] support for selecting metadata in single call expressions
• [Operations] fixed issue with using TypeNameHandling.None by user and deserialization of Operations
• [Operations] fixed issue with contacting server when using server-wide operations and node is a watcher

Studio

• [Licenses] display connectivity to api.ravendb.net
• [Licenses] show force update license on all license types
• [Notifications] added total documents size for page size is too big performance hint
• [Threads] added the ability to view a stack trace of a single thread

Server

• [Backups] fixed issue with restoring snapshots that were 2GB+ because of limitations of ZipArchive in .NET
• [Backups] fixed issue with Azure client having issues with staging large blocks
• [Backups] fixed progress update in AWS S3 and AWS Glacier clients
• [Cluster] decrease the amount of locks taken by cluster to avoid contention
• [Cluster] check for topology changes also on server-wide requests resulting in faster topology change discovery by Client APIs
• [Configuration] added 'Backup.UploadTimeoutInMin'
• [Indexes] fixed possible storage corruption in indexes that are using Suggestions
• [Indexes] fixed issue when modifying 'PatternReferencesCollectionName' in map-reduce indexes could create conflicts during replication
• [Indexes] fixed issue with using 'IfEntityIs'
• [Indexes] fixed NRE when database compaction is executed - could happen if auto indexes were merged
• [Indexes] prevent database from unloading because of idleness during index compaction
• [Linux] set 'TasksMax' to infinity in 'ravendb.service' file
• [Memory] fixed memory leak that could occur in Databases cache when a lot of databases is being added and deleted
• [Monitoring] exposed writes and data written per second in SNMP for server and per database
• [Monitoring] exposed average request duration for server
• [Querying] better exception when page size is negative
• [Replication] fixed issue that could cause replication problem if a revision of a conflict between document and document deletion occurs
• [Revisions] properly set etag when revision bin entry is created to avoid issues with ordering and therefore retrieving the revisions bin entries
• [Voron] immediate process exit on catastrophic failure in server store to avoid running with invalid state

Client API

• [Session] using 'Evict' method should release memory immediately to release memory pressure as soon as possible
• [Operations] fixed issue with deserialization of 'IOperationResult' when 'TypeNameHandling.None' is set in conventions
• [Operations] fixed possible NRE when waiting for smuggler operation if it was completed before the wait was called

Embedded

• added support for + in ServerOptions.FrameworkVersion for Patch number with default set to 3.1.19+

Tools

• [rvn] proper 'CTRL+D' and 'EOF' handling

Docker

• added support for $RAVEN_DATABASE parameter which creates given database on a first run
• added support for 'arm64' docker image

Server

• [Backups] moved to official Azure, AWS S3 and AWS Glacier clients
• [Concurrency] fixed possibility of loosing a document tombstone when optimistic concurrency 'put' is done after 'delete'
• [Configuration] added 'Storage.DiscardVirtualMemory' option
• [Encryption] fixed issue with removing encrypted node with wrong key
• [Indexing] decreased CPU pressure by scheduling indexing cleanup only when it is needed
• [Indexing] avoid creating index writer if no results were generated
• [Indexing] handle properly OperationCancelledException and ObjectDisposedException when querying an index that is being replaced at the same time
• [Indexing] handle 'SelectMany' in nested dictionary properly in indexing functions
• [Indexing] fixed select value from collection in indexing function while using 'FirstOrDefault' method
• [Monitoring] added endpoints to monitor number of indexes, stale indexes and indexing performance in all loaded databases
• [Querying] when projecting from a document that does not exist already, we should skip that document
• [Replication] when replication failovers it should check destination change vector and start from there, not from scratch
• [Revisions] enforcing revision configuration should also take into account document tombstones
• [Revisions] added 'MaximumRevisionsToDeleteUponDocumentUpdate' option
• [Voron] fixed issue with recursive write lock acquisitions when removing inactive scratches
• [Voron] fixed possibility of accessing disposed pager of 'compression.buffers'
• [Voron] fixed int.MaxValue overflow in 'AdjustPagesRequiredFor32Bits' when running on 32-bits
• [.NET Core] updated to 3.1.17

Client API

• [Querying] fixed issue with using 'Group' as a property name in includes
• [Session] method Attachmnts.GetNames should return correct number of attachments after deletion
• [Session] properly detect changes in documents containing Unicode characters
• [Subscriptions] added 'Disabled' to 'SubscriptionCreationOptions'

Studio

• [Performance Hints] added Total Documents Size to paging performance hint
• [Replication] added number of replicated counters to Replication Stats Graph

Tools

• [Debug] added ability to save traffic watch

Server

• [Backups] restore of compare exchange should consider compare exchange tombstones and clean them
• [Certificates] consider all of the certificates that match pinning hash to allow the connections after replacement
• [Cluster] fixed issue with huge Raft commands impacting stability of the cluster
• [Cluster] fixed issue with cluster transaction processing hang when server is restarted immediately after such a transaction occurs
• [Configuration] added 'Storage.MaxNumberOfRecyclableJournals'
• [Indexes] added support for ThenBy and ThenByDescending
• [Indexes] fixed possibility of a deadlock when max number of concurrent indexing processes is set
• [Indexes] fixed index instance leak when database record change is being handled but database is being unloaded at the same time
• [Licensing] try to activate license from environment variable or path when we fail to do it online
• [Memory] try to reuse large segments from the memory allocator (possible memory leak when doing long operations like backups or export)
• [Monitoring] added ability to define secondary user for SNMP v3
• [Replication] fixed possibility of replication killing server when database is being deleted at the same time
• [Replication] fixed possibility of replication with missing attachments went into endless loop state
• [Replication] fixed possible endless loop caused by not handling attachment stream files properly when destination database in encrypted
• [Replication] fixed possible NRE when Documents DeleteRevision flag was not stripped properly
• [Queries] fixed issue with delete or patch by query timing out when a lot of document IDs needs to be gathered
• [Voron] limiting number of recyclable journals to 32 by default
• [Voron] fixed case where a newly created read transaction could get an already disposed pager state of a scratch file
• [Utils] fixed issue with DisposeAsync not preventing double dispose in certain situations
• [.NET Core] updated to .NET Core 3.1.16

Client API

• [Bulk Insert] fixed possible hang and avoid multi-threaded access to context
• [JSON] properly clearing internal JsonReader state to avoid hitting MaxDepth exception

Studio

• [Migration] fixed issue with Advanced menu not opening
• [Setup] display unsupported browser warning if needed

Server

• [Backups] fixed issue with executing Backup Retention policies on Azure
• [Backups] allow to restore a document when collection was changed in the same incremental backup
• [Backups] fixed possible memory leak during import when a lot of compare exchange values are being imported
• [Debug] fixed issue with capturing stacktraces for entire cluster
• [Debug] gixed start date of a cloned transaction in /admin/debug/txinfo endpoint
• [Databases] fixed NRE taht could happen if database was created with Client Configuration set
• [Indexes] release temporary files on low memory to reduce the memory pressure further
• [Indexes] fixed a race between creating a new auto index and deleting the superseded one
• [Indexes] fixed usage of concurrent indexes semaphore to avoid NRE
• [Replication] fixed memory allocation when handling tombstones as a conflicts resolutions
• [Replication] fixed NRE on refresh of document that was added in cluster transaction
• [Voron] making sure that we will not remove currently used scratch buffer file from the list of active scratches during transaction rollback
• [Voron] making the acquirement and usage of a pager state more robust to avoid accessing already disposed pager which could result in AVE
• [Voron] preventing a racy access and dispose of a scrach file between the storage cleanup and creation of a new transaction
• [Voron] fixing the case that after the import of a database we didn't flush all journals if there was no activity on a database what resulted in large number of scratch files
• [Voron] using a correct pager state when prefetching segments
• [.NET Core] updated to .NET Core 3.1.15

Client API

• [Session] fixed caching when Lazy is used
• [Session] properly incrementing number of requests when attachments and revisions are used
• [Smuggler] fixed possibility to context leaking during Import
• [Querying] added support for aggregation queries when raw query is used

Studio

• [Indexes] expose index wait time on performance graph when max number of concurrent indexes is set
• [Indexes] fixed issue with displaying tooltips on performance graph
• [Refresh] fixed issue with updating the refresh frequency
• [Traffic Watch] display also seconds in request time
• [Traffic Watch] added TCP connections