Datadog

The trouble with mounting | Datadog

Jan 23, 2017Greg Meyer, Yann Mahe
The trouble with mounting | Datadog
SummaryAI, from the source post

This post details the investigation and resolution of Datadog Agent hangs caused by `os.statvfs` calls on NFS mounts. The core issue was identified as `statvfs` blocking indefinitely on hard-mounted NFS shares, especially when the NFS server was unresponsive or the connection was flaky. The solution implemented was to run the `os.statvfs` call on a separate thread with a timeout mechanism, allowing the main agent thread to continue processing even if the disk check hangs. This addresses the unkillable hang state observed in previous versions.