Dan Kinder

I just started up a fresh 256-node cluster on `V21.1.0-alpha.3` and found the CPU overhead of a totally idle cluster surprisingly high. Each node is consuming 200-300% CPU (about 23%...

I didn't compare to the previous, I've been using this release for the past month or two since I need some of the changes in it.

This includes a 5-second CPU profile, heap, and heap delta: [Archive.zip](https://github.com/cockroachdb/cockroach/files/6407893/Archive.zip)

We were just bit by this. We had a query object we were reusing, where the first use didn't read all the records and closed the iterator early, and then...

Yeah in my case it is not an issue with concurrent access. It does sound like the page state is simply not reset. Maybe fixing that would be quicker/easier than...

I just realized that there are 2 other jobs failing, so their failure could be related. One is an REFRESH MATERIALIZED VIEW, the other is "Migration to 21.2-56: "populate RangeAppliedState.RaftAppliedIndexTerm...

10. Most created a year ago or more.

I have a debug zip if you want it, thing is, there are 2 nodes that won't have logs included. One is the one that died last week and is...

Another node just died and won't start back up with this message: ``` ERROR: startup forbidden by prior critical alert DETAIL: From /srv/disk4/cockroach-data2/auxiliary/_CRITICAL_ALERT.txt: ATTENTION: this node is terminating because a...

This may help explain why that migration isn't finishing: ``` I220809 19:33:30.973834 96742 migration/migrationmanager/manager.go:372 ⋮ [n47,intExec=‹set-version›,migration-mgr] 109 found existing migration job 785222573203750976 for version 21.2-56 in status running, waiting I220809...