xklonx
commented
4 years ago +1 it happened once in my console application, none of the other tasks started, and every half an hour another worker was connected to this job.
I did not find anything in the method that could take so long and when I started the debug, none of the breakpoints worked.
I use DisableConcurrentExecutionAttribute, looks like for some reason distributed lock is stucked and when i restarted application it was released.
I think its easier to reproduce when multiple recurring jobs running every minute.
Happend again and i dont know why jobs not planned even when only two of them stucked
Looks like its caused by HangFire:recurring-jobs:lock that is not released for some reason, perhaps the reason is PostgreSqlStorageOptions.DistributedLockTimeout 24 hours that i use instead of default 10 minutes, which lead to several parallel tasks execution after 10 minutes.
The reason is this error, dont know what to do with it
Recurring job '***' can not be scheduled due to an exception.
System.Transactions.TransactionAbortedException: The transaction has aborted. ---> System.TimeoutException: Transaction Timeout
Example here https://github.com/frankhommers/Hangfire.PostgreSql/issues/119#issuecomment-605337122
hakancunier
RecurringJob.AddOrUpdate to Create Job。 Sometimes it happens that the job is always in processing state, but it seems that the job has been released. I've added a lot of logs to this code, but even if it's half recorded, it's gone, and there's no exception. If I trigger this job again, it runs normally.