Filebeat 5.6 error

[Samuel_Miller]

I'm seeing the error below, it looks extremely similar to
Error when reloading prospectors for filebeat - #7 by ruflin (Filebeat: Live Reload incorrectly stops prospectors · Issue #4133 · elastic/beats · GitHub) , which was apparently resolved, but I'm still seeing the error in Filebeat 5.6. Is there any way to resolve this?

2018-01-08T19:36:17Z ERR Error loading states for prospector 10998142644621256744: Can only start a prospector when all related states are finished: {Source:/logs/containers/ed498d049e5fdd17760af3d29e3cc14d4c3fb3f49816490ef4767fcfd3271854/ed498d049e5fdd17760af3d29e3cc14d4c3fb3f49816490ef4767fcfd3271854-json.log Offset:245013006 Finished:false Fileinfo:0xc42026f380 FileStateOS:{Inode:3821251 Device:51713} Timestamp:2018-01-08 19:35:42.264144712 +0000 UTC m=+305.083664000 TTL:-1ns}

[Samuel_Miller]

I've worked around this error by not running Filebeat from within a container. This is pretty disappointing, but I wasn't able to wait around for another fix to an apparently already resolved issue.

[ruflin]

@Samuel_Miller I just checked our 5.6 branch and it should be indeed fixed there. We even have a test in 5.6 that checks that the bug is resolved.

In case of reloading, the above error can still show up if you have 2 prospectors with overlapping configs which is not a bug. Could you share your config?

I wonder why running it inside the container solves your issue?