Hangfire job freezes after a few seconds of run in linux container

[jamespaulb]

We are using Hangfire.AspNetCore Version 1.8.14 and the hangfire.PostgreSql version 1.20.8 with .NET Core 8.0 in a linux container where we have multiple instances of application running.

The job starts and freezes after a few seconds. There are no exception or error messages shown.

[azygis]

We'll need more details and preferably a reproduction for this. We've been using Hangfire in linux container for years and plenty of versions, both .NET and Hangfire. Sounds more like the job logic itself hangs somewhere.

[jamespaulb]

If i use one instance of the application, then the job runs without issues. If i increase the number of instances of applications to 2 or more, then each instance processes a separate job. We need multiple instances as we need to process the different requests parallelly. Does this lead to any deadlock or concurrency issue?

[azygis]

Again, we need more details. What are the jobs doing? How is Hangfire set up?

We use 1-4 instances depending on environment. Never had a job just stuck, other than a bug in the job itself which ended up waiting indefinitely for a condition that never went true.

[jamespaulb]

Thank you azygis. The job reads a csv file consisting of millions of rows and updates the data of one or more columns in it. Are there any recommended versions of Hangfire.AspNetCore and hangfire.PostgreSql for .NET Core 8.0?

[azygis]

There is no particular recommended version. Any version works.

"Millions of rows" may incur OOM exception which nay or may not be logged, depending on whether the container stays alive for long enough, and also depending on whether it's all in memory or not. With such a vague job description I'm afraid I cannot help, still. This may as well be just infrastructure issue not related the the library.