got 2 log errors every minute in latest 1.10.x

[infofromca]

2016-04-15 18:45:09,214 [84] NHibernate.AssertionFailure - (null) - An AssertionFailure occurred - this may indicate a bug in NHibernate or in your custom types. [(null)]
NHibernate.AssertionFailure: null id in Orchard.Tasks.Locking.Records.DistributedLockRecord entry (don't flush the Session after an exception occurs)
2016-04-15 18:45:09,242 [84] Orchard.Tasks.Locking.Services.DistributedLockService - (null) - An error occurred while trying to acquire lock 'DistributedLock:RealEstateWorld:Orchard.JobsQueue.Services.JobsQueueProcessor'. [(null)]
NHibernate.AssertionFailure: null id in Orchard.Tasks.Locking.Records.DistributedLockRecord entry (don't flush the Session after an exception occurs)
   at NHibernate.Event.Default.DefaultFlushEntityEventListener.CheckId(Object obj, IEntityPersister persister, Object id, EntityMode entityMode)
   at NHibernate.Event.Default.DefaultFlushEntityEventListener.GetValues(Object entity, EntityEntry entry, EntityMode entityMode, Boolean mightBeDirty, ISessionImplementor session)
   at NHibernate.Event.Default.DefaultFlushEntityEventListener.OnFlushEntity(FlushEntityEvent event)
   at NHibernate.Event.Default.AbstractFlushingEventListener.FlushEntities(FlushEvent event)
   at NHibernate.Event.Default.AbstractFlushingEventListener.FlushEverythingToExecutions(FlushEvent event)
   at NHibernate.Event.Default.DefaultFlushEventListener.OnFlush(FlushEvent event)
   at NHibernate.Impl.SessionImpl.Flush()
   at NHibernate.Transaction.AdoTransaction.Commit()
   at Orchard.Data.TransactionManager.DisposeSession() in C:\Users\andy\Documents\orchard110house-zg\Orchard1203\Orchard\src\Orchard\Data\SessionLocator.cs:line 96
   at Orchard.Data.TransactionManager.Dispose() in C:\Users\andy\Documents\orchard110house-zg\Orchard1203\Orchard\src\Orchard\Data\SessionLocator.cs:line 86
   at Autofac.Core.Disposer.Dispose(Boolean disposing)
   at Autofac.Util.Disposable.Dispose()
   at Autofac.Core.Lifetime.LifetimeScope.Dispose(Boolean disposing)
   at Autofac.Util.Disposable.Dispose()
   at Orchard.Tasks.Locking.Services.DistributedLockService.ExecuteOnSeparateTransaction(Action`1 action) in C:\Users\andy\Documents\orchard110house-zg\Orchard1203\Orchard\src\Orchard\Tasks\Locking\Services\DistributedLockService.cs:line 215
   at Orchard.Tasks.Locking.Services.DistributedLockService.EnsureDistributedLockRecord(String internalName, Nullable`1 maxValidFor) in C:\Users\andy\Documents\orchard110house-zg\Orchard1203\Orchard\src\Orchard\Tasks\Locking\Services\DistributedLockService.cs:line 142
   at Orchard.Tasks.Locking.Services.DistributedLockService.<>c__DisplayClass9_1.<AcquireLockInternal>b__0() in C:\Users\andy\Documents\orchard110house-zg\Orchard1203\Orchard\src\Orchard\Tasks\Locking\Services\DistributedLockService.cs:line 103
   at Orchard.Tasks.Locking.Services.DistributedLockService.RepeatUntilTimeout(Nullable`1 timeout, TimeSpan repeatInterval, Func`1 action) in C:\Users\andy\Documents\orchard110house-zg\Orchard1203\Orchard\src\Orchard\Tasks\Locking\Services\DistributedLockService.cs:line 198
   at Orchard.Tasks.Locking.Services.DistributedLockService.AcquireLockInternal(String name, Nullable`1 maxValidFor, Nullable`1 timeout, Boolean throwOnTimeout) in C:\Users\andy\Documents\orchard110house-zg\Orchard1203\Orchard\src\Orchard\Tasks\Locking\Services\DistributedLockService.cs:line 102
2016-04-15 18:45:09,263 [84] Orchard.Tasks.Locking.Services.DistributedLockService - (null) - Error while trying to acquire lock 'Orchard.JobsQueue.Services.JobsQueueProcessor'. [(null)]
NHibernate.AssertionFailure: null id in Orchard.Tasks.Locking.Records.DistributedLockRecord entry (don't flush the Session after an exception occurs)
   at NHibernate.Event.Default.DefaultFlushEntityEventListener.CheckId(Object obj, IEntityPersister persister, Object id, EntityMode entityMode)
   at NHibernate.Event.Default.DefaultFlushEntityEventListener.GetValues(Object entity, EntityEntry entry, EntityMode entityMode, Boolean mightBeDirty, ISessionImplementor session)
   at NHibernate.Event.Default.DefaultFlushEntityEventListener.OnFlushEntity(FlushEntityEvent event)
   at NHibernate.Event.Default.AbstractFlushingEventListener.FlushEntities(FlushEvent event)
   at NHibernate.Event.Default.AbstractFlushingEventListener.FlushEverythingToExecutions(FlushEvent event)
   at NHibernate.Event.Default.DefaultFlushEventListener.OnFlush(FlushEvent event)
   at NHibernate.Impl.SessionImpl.Flush()
   at NHibernate.Transaction.AdoTransaction.Commit()
   at Orchard.Data.TransactionManager.DisposeSession() in C:\Users\andy\Documents\orchard110house-zg\Orchard1203\Orchard\src\Orchard\Data\SessionLocator.cs:line 96
   at Orchard.Data.TransactionManager.Dispose() in C:\Users\andy\Documents\orchard110house-zg\Orchard1203\Orchard\src\Orchard\Data\SessionLocator.cs:line 86
   at Autofac.Core.Disposer.Dispose(Boolean disposing)
   at Autofac.Util.Disposable.Dispose()
   at Autofac.Core.Lifetime.LifetimeScope.Dispose(Boolean disposing)
   at Autofac.Util.Disposable.Dispose()
   at Orchard.Tasks.Locking.Services.DistributedLockService.ExecuteOnSeparateTransaction(Action`1 action) in C:\Users\andy\Documents\orchard110house-zg\Orchard1203\Orchard\src\Orchard\Tasks\Locking\Services\DistributedLockService.cs:line 215
   at Orchard.Tasks.Locking.Services.DistributedLockService.EnsureDistributedLockRecord(String internalName, Nullable`1 maxValidFor) in C:\Users\andy\Documents\orchard110house-zg\Orchard1203\Orchard\src\Orchard\Tasks\Locking\Services\DistributedLockService.cs:line 142
   at Orchard.Tasks.Locking.Services.DistributedLockService.<>c__DisplayClass9_1.<AcquireLockInternal>b__0() in C:\Users\andy\Documents\orchard110house-zg\Orchard1203\Orchard\src\Orchard\Tasks\Locking\Services\DistributedLockService.cs:line 103
   at Orchard.Tasks.Locking.Services.DistributedLockService.RepeatUntilTimeout(Nullable`1 timeout, TimeSpan repeatInterval, Func`1 action) in C:\Users\andy\Documents\orchard110house-zg\Orchard1203\Orchard\src\Orchard\Tasks\Locking\Services\DistributedLockService.cs:line 198
   at Orchard.Tasks.Locking.Services.DistributedLockService.AcquireLockInternal(String name, Nullable`1 maxValidFor, Nullable`1 timeout, Boolean throwOnTimeout) in C:\Users\andy\Documents\orchard110house-zg\Orchard1203\Orchard\src\Orchard\Tasks\Locking\Services\DistributedLockService.cs:line 134
   at Orchard.Tasks.Locking.Services.DistributedLockService.TryAcquireLock(String name, Nullable`1 maxValidFor, Nullable`1 timeout, IDistributedLock& dLock) in C:\Users\andy\Documents\orchard110house-zg\Orchard1203\Orchard\src\Orchard\Tasks\Locking\Services\DistributedLockService.cs:line 42

[jtkech]

Which database do you use?

See #6729, only applied to the dev branch, where it was needed to do the RequireNew() before creating the schema of the DistributedLockRecord table, this when using a PostgreSQL database. Maybe you have the same issue in 1.10.x.

But, we must be careful, it seems that we still need (at least with sqlce & sql server) to do a RequireNew() after the schema creation, see #6783.

Best.

[infofromca]

compact ce

[sebastienros]

Let's revert the fix for now if it's creating another issue, and reopen the initial bug

[sebastienros]

So I am lost, the PostreSQL fix is only on dev, can everyone repro the issue on 1.10.x ? Since which commit then?

[infofromca]

IT was NOT good ON 1.10 RELEASE

[jtkech]

@sebastienros

So, i think the issue of @infofromca is on 1.10 but before this fix #6737 by @sfmskywalker. So, not related with #6729 only applied to dev.

I mentioned #6729 in case @infofromca would use e.g a PostgreSQL database.

[infofromca]

it has been from 1.10 to 1.10.x today. which means not relate with #6737
I tested that in two.
got ex inside ExecuteOnSeparateTransaction method.
this is 1.10

this is today's 1.10.x

[jtkech]

@sebastienros @infofromca

Oh yes, i think i understand and i can simulate the use case in last 1.10.x.

So, it works as long as a lock record is acquired then deleted after the related job ended. But if a lock record is not deleted this way (this is my simulation), the following code added in #6737 will try to remove it before creating a new one:

        foreach (var expiredRecord in records) {
            repository.Delete(expiredRecord);
        }

        // No valid record existed, so we're good to create a new one.
        Logger.Debug("No valid record was found for lock '{0}'; creating a new record.", ...);

        repository.Create(new DistributedLockRecord {
        ...

So, if you already have a lock record, i think the issue is that we can't delete a lock record and create a new one with the same index in the same transaction.

Best

[jtkech]

@sebastienros

Then, because we are already in a new transaction, so i think we can add a flush after the delete. Then, it seems to work.

                    repository.Delete(expiredRecord);
                    // Evolutive
                    repository.Flush();

Best.

[infofromca]

@jtkech @sebastienros
I also added repository.Flush(); inside the foreach, verified that it is ok

[sebastienros]

Even though Flush is usually not the solution, here it is correct.

[sebastienros]

PR please

[jtkech]

If @infofromca want to do it, otherwise i can do it.

[jtkech]

@sebastienros

Another solution would be to use another transaction as in the DeleteDistributedLockRecord() method, but maybe too many transactions.

Or just an idea on the fly: not delete the record but only update the valid time.

Best.

[infofromca]

so this give me a idea:
Can we create AddOrUpdate method on IRepository<> ?