from kvctl-server we can find two type log for connectivity check :
1.Failed to ping the node,means kvrocks instance or node is down,this log has failure_count number。
2.i/o timeout,means connect to kvrocks instance is timeout ,this log no failure_count number。
for example:
{"level":"warn","ts":1691306009.5305622,"caller":"probe/cluster.go:107","msg":"Failed to ping the node","id":"tTN3gDB4lDemUvYMAg1223qJpaK5","role":"slave","addr":"192.1.151:6012","error":"EOF","failure_count":1}
{"level":"error","timestamp":"2024-04-12T15:11:16.209+0800","caller":"probe/cluster.go:125","msg":"Failed to probe the node","id":"pEmo25f6VPusukxKonHzCDeQ6ANNludNAzccSZ3k","role":"master",
"addr":"192.1.151:6011","error":"i/o timeout","stacktrace":"github.com/apache/kvrocks-controller/controller/probe.(*Cluster).probe.func1\n\t/opt/kvrocks_contoller/kvrocks-controller/cont
roller/probe/cluster.go:125"}
i think logs for “i/o timeout” also need failure_count number for read
from kvctl-server we can find two type log for connectivity check : 1.Failed to ping the node,means kvrocks instance or node is down,this log has failure_count number。 2.i/o timeout,means connect to kvrocks instance is timeout ,this log no failure_count number。
for example: {"level":"warn","ts":1691306009.5305622,"caller":"probe/cluster.go:107","msg":"Failed to ping the node","id":"tTN3gDB4lDemUvYMAg1223qJpaK5","role":"slave","addr":"192.1.151:6012","error":"EOF","failure_count":1}
{"level":"error","timestamp":"2024-04-12T15:11:16.209+0800","caller":"probe/cluster.go:125","msg":"Failed to probe the node","id":"pEmo25f6VPusukxKonHzCDeQ6ANNludNAzccSZ3k","role":"master", "addr":"192.1.151:6011","error":"i/o timeout","stacktrace":"github.com/apache/kvrocks-controller/controller/probe.(*Cluster).probe.func1\n\t/opt/kvrocks_contoller/kvrocks-controller/cont roller/probe/cluster.go:125"}
i think logs for “i/o timeout” also need failure_count number for read