HangFire.SqlServer 1.6.21

SQL Server 2008+ (including Express), SQL Server LocalDB and SQL Azure storage support for Hangfire (background job system for ASP.NET applications).

Release Notes

https://www.hangfire.io/blog/
   
1.6.21
• Fixed – Change locking scheme in the `Connection.SetRangeInHash` method to avoid deadlocks.
     
1.6.20
• Fixed – Remove the synthetic limitation to support Azure SQL Management instance (by @TimSQL).
     
1.6.19
• Fixed – "String or binary data would be truncated" exception when state reason is too long.
• Fixed – Command handling in batch mode now stops after the first error due to "XACT_ABORT ON".
• Fixed – Make SqlCommandBatch disposable to dispose all the commands.
     
1.6.18
• Fixed – Rare deadlocks in SQL Server caused by the SetRangeInHash command.
   
1.6.17
• Added – `SqlServerStorageOptions.CommandBatchMaxTimeout` parameter to enable batching in transactions.
• Fixed – Timeout exceptions when there are a lot of large concurrent transactions by using the new batching method.
• Fixed – Distributed locks are safe now even in very unreliable networks and after network blips.
• Performance – Greatly decreased the number of connections required to process background jobs.
• Performance – Significantly decrease the number of roundtrips required to commit a transaction.
   
1.6.16
• Fixed – Lower the number of requests, when trying to acquire a distributed lock.
• Fixed – Problems with continuations and batch continuations, when using existing `SqlConnection` instance.
• Fixed – Distributed locks were silently released, when passing explicit closed connection to a storage.
• Fixed – `KeyNotFoundException` errors on various dashboard pages don't appear anymore (by @benjymous).
• Fixed – "Cannot release the application lock because it is not currently held" exceptions, when connection was closed.
• Fixed – Allow to use zero timeout for SQL Server-based distributed locks.
   
1.6.15
• Fixed – Timeout exceptions that's caused by SQL Server's thread pool starvation, caused by sp_getapplock.
• Fixed – Antecedent background job is constantly failing, when its continuation hasn't been fully created.
   
1.6.14
• Added – `SqlServerStorageOptions.SlidingInvisibilityTimeout` to fetch jobs without using transaction.
• Fixed – Transaction log is full due to 'ACTIVE_TRANSACTION' by enabling to use new non-transactional fetch.
• Fixed – `SqlServerJobQueueMonitoringApi` can't cause READ UNCOMMITTED isolation level to leak on SQL Server 2012 or earlier.
• Fixed – Add missing `SqlServerStorage(DbConnection, SqlServerStorageOptions)` constructor.
   
1.6.13
• Fixed – Don't hide errors occurred while running SQL migrations.
• Fixed – `KeyNotFoundException` when accessing Deleted Jobs page in Dashboard.
• Fixed – `SqlServerDistributedLock` leaks connections, when lock acquisition is failed.
   
1.6.9
• Fixed – `TimeoutException` on large arguments or large batches via `SqlServerOptions.CommandTimeout`.

1.6.8
• Fixed – Use `long` where possible instead of `int` for background job identifiers, full support will be in 1.7.0.
   
1.6.7
• Fixed – Validation added to avoid "An invalid application lock time-out" exceptions (by t0mburton).
   
1.6.6
• Fixed – SqlServerJobQueue class doesn't use obsolete InvisibilityTimeout parameter anymore.
   
1.6.5
• Fixed – Different timeout issues after making performance optimizations.
• Fixed – SQL timeouts while getting a connection, when using default settings and >= 8 CPU cores.
• Fixed – ExpirationManager is bloated by SqlError instances, when there are a lot of messages from server.
• Fixed – Counters query returned inconsistent results during CountersAggregator executions.
• Performance – Added missing NOLOCK hint for monitoring queries when using SqlServer-based queues.
• Performance – ExpirationManager is forced to use index seek operations for cascade deletions.
• Performance – CountersAggregator now uses clustered index scan to issue less logical reads.
• Performance – Paging queries in dashboard now use CTEs to utilize less logical reads.
   
1.6.4
• Fixed – `SqlConnection` object leaks when using nested distributed locks (in batches and continuations).

1.6.3
• Fixed – Prevent `ExpirationManager`'s command from timing out, when there are a lot of records.
• Fixed – `CountersAggregator` issue introduced in 1.6.2, that causes slow record aggregation.
• Fixed – Add `XLOCK` hint for `CountersAggregator` query to prevent deadlocks on lock escalation.
   
1.6.2
• Changed – `JobExpirationCheckInterval` property value now defaults to 30 minutes.
• Fixed – Excessive blocking when expiring records in `State` and `JobParameters` tables.
• Fixed – Migration script `Install.sql` was deadlocked by itself, leading to errors on startup.

1.6.0
• Added – .NET Core support for Hangfire.Core and Hangfire.SqlServer packages (`netstandard1.3` target).
• Added – Zero latency processing when client and server reside in the same process (by @bgurmendi).
• Changed – `SqlServerStorage` class constructor now takes `DbConnection` instances (by @tystol).
• Changed – `SqlServerObjectsInstaller` class is public now (by @vbfox).
• Changed – Update Dapper to 1.50.1 version.
• Fixed – Drop constraints on Server.Id prior to altering it to avoid conflicts.

Dependencies

• .NETFramework 4.5

  • Hangfire.Core (= 1.6.21)

• .NETStandard 1.3

  • Dapper (>= 1.50.1)
  • Hangfire.Core (= 1.6.21)
  • Microsoft.NETCore.Portable.Compatibility (>= 1.0.1)
  • NETStandard.Library (>= 1.6.0)
  • System.Data.Common (>= 4.1.0)
  • System.Data.SqlClient (>= 4.1.0)

Version History