gc_grace_seconds set is the default 10 days (these tables do not have data with ttl, or anything else that might create tombstones), and max_hint_window_in_ms parameter is set to 18h. I'm checking it by looking at both the hints directory and hints metrics, and as far as check time, I took a node down for an hour and there are plenty of writes to cause a flush, but I will next time issue a flush manually. I checked 2 nodes so far out of dozens, my next step is to check on more, the application should be using them all, could be that's not the case.
On Fri, May 28, 2021 at 2:18 AM Kane Wilson <k...@raft.so> wrote: > > Hey >> I have a cluster, 3.11.9, in which I am enabling hints, but it appears >> that no hints are created when other nodes are down. >> I can see that hints are enabled by running "nodetool statushandoff", and >> gc_grace_seconds is high enough on all tables, so I'm expecting to see >> hints being created, since there are writes constantly coming, but nothing >> is being created. >> Any ideas? >> Gil >> > > Hints are created on the coordinator. Have you checked all nodes? Maybe > they haven't been flushed to disk yet. > Also what is "high enough" in relation to GCGS? If it's less than the > configured hinted handoff duration they'll be dropped. Are you checking the > relevant hints metrics? They will give you a good indication if hints are > actually occurring. > > -- > raft.so - Cassandra consulting, support, and managed services >
