mds: fragmentation in QA is slowing down ops enough for WRNs
Here's a more obvious snippet from the logs:
2018-05-09 16:36:21.577 7f12cf1c4700 0 log_channel(cluster) log [WRN] : 1 slow requests, 1 included below; oldest blocked for > 34.482641 secs 2018-05-09 16:36:21.577 7f12cf1c4700 0 log_channel(cluster) log [WRN] : slow request 34.482640 seconds old, received at 2018-05-09 16:35:47.100911: internal op fragmentdir:mds.0:1 currently submit entry: dispatch_fragment_dir
I'm not really happy whitelisting these types of warnings but we also need to quiet this warning. Any thoughts?