Pipelines stop with Java::OrgJrubyExceptions::SystemCallError

[manuel.ferreira]

Hi.

Some light on this issue would be very much appreciated. The scenario is the following.

Logstash pipelines fail after docker restart.
Logstash : 7.9.3

• When logstash exits with an error, the docker configuration restarts the container.
• Upon restart we start getting "exception": "No such file or directory on /tmp files
• message": "The XML you provided was not well-formed or did not validate against our published schema", "path": "/tmp/
• "message": "An error occurred in the on_complete uploader", "exception": { "metaClass": { "metaClass": { "exception": "Errno::ENOENT", "message": "No such file or directory - /tmp/
• "message": "File doesn't exist! Unrecoverable error.", "exception": { "metaClass": { "metaClass": { "exception": "Errno::ENOENT", "message": "No such file or directory - No such file or directory - /tmp/
• { "level": "WARN", "loggerName": "logstash.inputs.s3", "timeMillis": 1615901713503, "thread": "[tv]<s3", "logEvent": { "message": "Logstash S3 input, stop reading in the middle of the file, we will read it again when logstash is started" } }
• Eventually the pipelines fail.

I've only put some mesages here for ease of reading but i'm happy to provide further detail.

It seems that upon the container restart, logstash retains state and when it's going for temp files, they don't exist anymore.

Anyone stumbled upon this?
Is this expected behavior? Is it possible to bypass?

Thanks in advance,
Manuel