Source: redis
Version: 4:4.0.1-6
Severity: serious
https://buildd.debian.org/status/package.php?p=redis&suite=sid
...
Cleanup: may take some time... OK
taskset --cpu-list 0 ./runtest-sentinel
Starting sentinel #0 at port 20000
Starting sentinel #1 at port 20001
Starting sentinel #2 at port 20002
Starting sentinel #3 at port 20003
Starting sentinel #4 at port 20004
Starting redis #0 at port 30000
Starting redis #1 at port 30001
Starting redis #2 at port 30002
Starting redis #3 at port 30003
Starting redis #4 at port 30004
Testing unit: 00-base.tcl
05:27:25> (init) Restart killed instances: OK
05:27:25> (init) Remove old master entry from sentinels: OK
05:27:25> (init) Create a master-slaves cluster of 5 instances: OK
05:27:26> (init) Sentinels can start monitoring a master: OK
05:27:26> (init) Sentinels can talk with the master: OK
05:27:26> (init) Sentinels are able to auto-discover other sentinels: OK
05:27:28> (init) Sentinels are able to auto-discover slaves: OK
05:27:28> Basic failover works if the master is down: OK
05:27:31> New master 127.0.0.1:30002 role matches: OK
05:27:31> All the other slaves now point to the new master: OK
05:27:31> The old master eventually gets reconfigured as a slave: OK
05:27:42> ODOWN is not possible without N (quorum) Sentinels reports: OK
05:27:42> Failover is not possible without majority agreement: OK
05:27:43> Failover works if we configure for absolute agreement: OK
05:27:47> New master 127.0.0.1:30000 role matches: OK
Testing unit: 01-conf-update.tcl
05:27:47> (init) Restart killed instances: OK
05:27:47> (init) Remove old master entry from sentinels: OK
05:27:47> (init) Create a master-slaves cluster of 5 instances: OK
05:27:48> (init) Sentinels can start monitoring a master: OK
05:27:48> (init) Sentinels can talk with the master: OK
05:27:48> (init) Sentinels are able to auto-discover other sentinels: OK
05:27:50> (init) Sentinels are able to auto-discover slaves: OK
05:27:50> After Sentinel 1 is restarted, its config gets updated: OK
05:27:50> New master 127.0.0.1:30000 role matches: OK
Testing unit: 02-slaves-reconf.tcl
05:27:50> (init) Restart killed instances: OK
05:27:50> (init) Remove old master entry from sentinels: OK
05:27:50> (init) Create a master-slaves cluster of 5 instances: OK
05:27:50> (init) Sentinels can start monitoring a master: OK
05:27:50> (init) Sentinels can talk with the master: OK
05:27:50> (init) Sentinels are able to auto-discover other sentinels: OK
05:27:52> (init) Sentinels are able to auto-discover slaves: OK
05:27:52> Check that slaves replicate from current master: OK
05:27:52> Crash the master and force a failover: OK
05:27:56> Check that slaves replicate from current master: OK
05:28:06> Kill a slave instance: OK
05:28:06> Crash the master and force a failover: OK
05:28:09> Check that slaves replicate from current master: OK
05:28:19> Wait for failover to end: OK
05:28:19> Restart killed slave and test replication of slaves again...: OK
05:28:19> Check that slaves replicate from current master: OK
Testing unit: 03-runtime-reconf.tcl
Testing unit: 04-slave-selection.tcl
Testing unit: 05-manual.tcl
05:28:40> (init) Restart killed instances: OK
05:28:40> (init) Remove old master entry from sentinels: OK
05:28:40> (init) Create a master-slaves cluster of 5 instances: OK
05:28:41> (init) Sentinels can start monitoring a master: OK
05:28:41> (init) Sentinels can talk with the master: OK
05:28:41> (init) Sentinels are able to auto-discover other sentinels: OK
05:28:43> (init) Sentinels are able to auto-discover slaves: OK
05:28:43> Manual failover works: OK
05:28:45> New master 127.0.0.1:30004 role matches: OK
05:28:45> All the other slaves now point to the new master: OK
05:28:45> The old master eventually gets reconfigured as a slave: OK
Testing unit: 06-ckquorum.tcl
05:28:55> (init) Restart killed instances: OK
05:28:55> (init) Remove old master entry from sentinels: OK
05:28:55> (init) Create a master-slaves cluster of 5 instances: OK
05:28:55> (init) Sentinels can start monitoring a master: OK
05:28:55> (init) Sentinels can talk with the master: OK
05:28:55> (init) Sentinels are able to auto-discover other sentinels: OK
05:28:57> (init) Sentinels are able to auto-discover slaves: OK
05:28:57> CKQUORUM reports OK and the right amount of Sentinels: OK
05:28:57> CKQUORUM detects quorum cannot be reached: OK
05:28:57> CKQUORUM detects failover authorization cannot be reached: kill:
(26442): No such process
while executing
"exec kill -9 $pid"
(procedure "kill_instance" line 9)
invoked from within
"kill_instance sentinel 1"
("uplevel" body line 4)
invoked from within
"uplevel 1 $code"
(procedure "test" line 6)
invoked from within
"test "CKQUORUM detects failover authorization cannot be reached" {
set orig_quorum [expr {$num_sentinels/2+1}]
S 0 SENTINEL SET mymaster quoru..."
(file "../tests/06-ckquorum.tcl" line 20)
invoked from within
"source $test"
(procedure "run_tests" line 9)
invoked from within
"run_tests"
(procedure "main" line 5)
invoked from within
"main"
Cleaning up...
debian/rules:30: recipe for target 'override_dh_auto_test' failed
make[1]: *** [override_dh_auto_test] Error 1
