File: gr_recovery_donor_failover_mysql.test

package info (click to toggle)
mysql-8.0 8.0.43-3
  • links: PTS, VCS
  • area: main
  • in suites: sid
  • size: 1,273,924 kB
  • sloc: cpp: 4,684,605; ansic: 412,450; pascal: 108,398; java: 83,641; perl: 30,221; cs: 27,067; sql: 26,594; sh: 24,181; python: 21,816; yacc: 17,169; php: 11,522; xml: 7,388; javascript: 7,076; makefile: 2,194; lex: 1,075; awk: 670; asm: 520; objc: 183; ruby: 97; lisp: 86
file content (22 lines) | stat: -rw-r--r-- 1,114 bytes parent folder | download | duplicates (2) 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
 
################################################################################
# Recovery donor failover test
# This test focus on the failover mechanism where a joiner changes it's donor
# whenever the one in use leaves the group while recovery is happening.
#
# To test this, a new member is started but recovery blocked. While at this state
# the donor is killed, forcing the joiner to change to a new one. As there is
# no donor identification as of now, we just unblock and check that recovery
# was successful
#
# Test:
# 0. The test requires three servers: M1, M2 and M3.
# 1. Bootstrap group on M1. Insert some data. Start GR on M2.
# 2. Stop the applier on both servers to block recovery.
# 3. Start GR on M3. It should go into RECOVERING state. Identify the DONOR.
# 4. Stop GR on the donor making it leave the group.
# 5. Unblock recovery i.e. Start the applier on failover member and watch M3 go
#    ONLINE despite the donor exit. Vaidate data on M3.
# 6. Clean up.
################################################################################

--source ../include/gr_recovery_donor_failover.inc