HangFire.SqlServer 1.7.14

• .NETFramework 4.5

  • Hangfire.Core (= 1.7.14)

• .NETStandard 1.3

  • Hangfire.Core (= 1.7.14)
  • NETStandard.Library (>= 1.6.0)
  • System.Data.Common (>= 4.1.0)
  • System.Data.SqlClient (>= 4.1.0)

• .NETStandard 2.0

  • Hangfire.Core (= 1.7.14)
  • System.Data.SqlClient (>= 4.4.0)

https://www.hangfire.io/blog/

Please see https://docs.hangfire.io/en/latest/upgrade-guides/upgrading-to-hangfire-1.7.html to learn how to upgrade from 1.6.X.

1.7.14
• Fixed – Duplicate entries in the `JobParameters` table after upgrading to version 1.7.13.
• Fixed – Extensive retries on a method that has a retry attribute after upgrading to 1.7.13.
• Fixed – "ArgumentException: An item with the same key has already been added. Key: RetryCount" in `SqlServerMonitoringApi`.

1.7.13
• Added – `UseIgnoreDupKeyOption` for SQL Server storage configuration (changes to [Set] and [Hash] tables required).
• Fixed – Don't truncate too long keys silently, throw exceptions instead.
• Fixed – Add missing null checks for methods in the `SqlServerWriteOnlyTransaction` class.
• Fixed – Change `holdlock` hint to `xlock` in `merge` statements in transaction to prevent deadlocks.
• Fixed – Don't rethrow "Lock request time out period exceeded" exceptions from expiration manager.
• Fixed – Increase [Server].[Id] column's length to 200 for new installations.

1.7.10
• Changed – Use `XACT_ABORT` option for `ExpirationManager` queries.
• Changed – Don't rely on implicit rollback when disposing transactions.
• Fixed – `NullReferenceException` in logs thrown from `SqlInternalTransaction.Rollback` method.

1.7.9
• Changed – Implement long polling fetch for sub-second polling delays without `sp_getapplock`.
• Fixed – Don't leak `DbConnection` instance when an exception occurs during its opening.
• Fixed – Can not obtain connection from the pool exception after database was offline.
• Fixed – High number of waits in SQL Server when Hangfire Servers are idle.

1.7.8
• Added – Support for Microsoft.Data.SqlClient package when using a custom connection factory (Part II).
• Fixed – Remove `System.Data.SqlClient` dependency from `SqlCommandBatch` and `ExpirationManager`.

1.7.7
• Added – Add support for Microsoft.Data.SqlClient package when using custom connection factory (Part I).
• Added – Add `UseFineGrainedLocks` option to avoid deadlocks in some theoretical cases.
• Added – Add missing overload for `UseSqlServerStorage` with connection factory parameter only.
• Added – Expose the SqlServerObjectsInstaller.GetInstallScript method (by @altso).
• Fixed – Make command batching working on .NET Core when using System.Data.SqlClient 4.7.0 and higher.
• Fixed – Permit dash characters (`-`) in schema names (by @kendaleiv).
• Fixed – Escape square bracket characters in schema names.
• Project – Add support for `netcoreapp3.0` target in Hangfire.SqlServer.Tests.
• Project – Take schema name from constant in Hangfire.SqlServer.Tests (by @kendaleiv).
• Project – Make Hangfire.SqlServer.Tests work on Linux in Travis CI environment.

1.7.4
• Fixed – Potential deadlocks cause by suboptimal queries when using `SlidingInvisibilityTimeout` fetching.
• Fixed – Prevent zero delays between fetch retry attempts when lock acquisition failed without blocking.
• Fixed – Specify float precision explicitly for the `Score` column in the `AddToSet` method.

1.7.3
• Fixed – Wrong error message in migration script, when @CurrentSchemaVersion has a NULL value (by @penenkel).

1.7.2
• Fixed – Occasional "DataException: Error parsing column" error when using blocking fetch.

1.7.1
• Changed – Use blocking fetch implementation only for sub-second polling intervals.
• Fixed – Don't fail with an exception when can't connect to MSSQL instance during start-up.
• Fixed – Don't access the `JobQueue` table when using blocking query and don't have results.

1.7.0
• Added – Full .NET Core 2.0 support by explicitly targeting .NET Standard 2.0.
• Added – `Schema 7` migration to fix the `IX_HangFire_Set_Score` index to include the `Key` column.
• Added – `Schema 6` migration with less indexes, better physical layout and `bigint` support (disabled by default).
• Added – Blocking fetch support for sliding expiration-based fetch to avoid excessive polling.
• Added – `SqlServerStorageOptions.EnableHeavyMigrations` switch to automatically install even heavy migrations.
• Added – `SqlServerStorageOptions.DisableGlobalLocks` property to avoid custom locking scheme.
• Added – `SqlServerStorageOptions.UsePageLocksOnDequeue` property to use less CPU consuming fetch.
• Added – Callback method to allow to open the database with impersonation (by @BjoernHund).
• Added – `SqlServerStorageOptions.UseRecommendedIsolationLevel` option to set the minimum possible level.
• Changed – Identity columns either converted to the `bigint` type, or entirely removed.
• Changed – Clustered indexes were organized according to the access patterns of their tables.
• Changed – Most of secondary indexes were either removed or made filtered.
• Changed – Optimize sliding-expiration-based fetching to use even less CPU time.
• Changed – Use write reordering and fine-grained locking scheme to improve parallelism.
• Changed – Monitoring API doesn't check state data to see state transition time.
• Changed – Allow to use zero-based poll interval when sliding invisibility timeout.
• Changed – Short paths for the `CreateExpiredJob` method to avoid some round-trips.
• Changed – Set `SqlParameter` types explicitly to not to duplicate query plans.
• Changed – Batch support for `AddToQueue` method when default provider is used.
• Changed – Check `FetchedAt` has expected value to prevent prolonging others' work.
• Changed – Use more recent Dapper 1.50.7 on all platforms except .NET Framework 4.5.
• Changed – Dapper package is internalized now even on .NET Core to avoid possible conflicts.
• Fixed – Avoid unnecessary waits in state changer when job was already expired.
• Fixed – Cannot resolve the collation conflict in `CountersAggregator`.
• Fixed – Background processing stops when identity columns exceed the `Int32.MaxValue`.
• Fixed – Slowdown of scheduled jobs due to the missing index on the `[Set]` table.