Description
Had 3 node cluster, started from cluster_run started much earlier in the day. Has 1 bucket with 1.3M items, and FTS index created on it.
It was completely idle, I forgot it was running and was browsing Amazon when I heard the cpu fans. Looking at top I realized all 3 cbft processes were consuming 100% cpu each.
I have gathered cpu profiles of each, but they show nothing. I killed one process with SIGABRT so I could get the goroutine dump, that too doesn't seem to show anything...
Attachments
Issue Links
- relates to
-
MB-19148 [FTS] forestdb auto-compaction is 8 hr cycle?
- Closed