Open dhbaird opened 4 years ago
Would be great to at least put a note in our docs about this, cc @ddorian @stevebang
Perhaps in some of the new core db best practices section?
@bmatican Pull request for docs: https://github.com/yugabyte/yugabyte-db/pull/4159
Docs are updated. Leaving open for feature request.
Jira Link: DB-1633 This ticket is more informational than an actual issue, because the underlying issue seemed to be a consequence of running on ZFS (zfsonlinux 0.8.3). And I should know better than that; But I wanted to document my experience in case it is helpful. I was able to easily workaround this issue by switching to another filesystem (ext4, though it should be noted that xfs is also recommended).
Yugabyte version: yugabyte-2.1.2.0
Without further ado: I was doing some very simple stress resiliency stress testing of trying to start yugabyte, and then kill it, and see how it reacts, of as follows:
Step 1: Install
Step 2: Try to break something (expectation: this shouldn't actually break anything):
I ran into trouble almost immediately:
(wait for a few minutes)
Then, kill node 1 processes and see what happens:
kill 3017 3008 Now I can't restart node.
~/yugabyte-data/node-1/disk-1/yb-data/master/logs/yb-master.myhost.personlog.INFO.20200406-190809.29819