-
Type:
Bug
-
Resolution: Duplicate
-
Priority:
Minor - P4
-
None
-
Affects Version/s: None
-
Component/s: Sharding
-
None
-
Service Arch
-
ALL
Before the ping time is explicitly set, the ping time is apparently max(int64). This gets reported badly via (shard)connPoolStats.
Example:
{
"hosts" : {
"nuwen:29000::0" : {
"available" : 1,
"created" : 1
},
"nuwen:29000::30" : {
"available" : 2,
"created" : 3
},
"nuwen:31100::0" : {
"available" : 0,
"created" : 1
},
"nuwen:31100::5" : {
"available" : 1,
"created" : 1
},
"test-rs0/nuwen:31100,nuwen:31101,nuwen:31102::0" : {
"available" : 1,
"created" : 1
}
},
"replicaSets" : {
"test-rs0" : {
"hosts" : [
{
"addr" : "nuwen:31100",
"ok" : true,
"ismaster" : true,
"hidden" : false,
"secondary" : false,
"pingTimeMillis" : 0
},
{
"addr" : "nuwen:31101",
"ok" : false,
"ismaster" : false,
"hidden" : false,
"secondary" : false,
"pingTimeMillis" : -1511828489
},
{
"addr" : "nuwen:31102",
"ok" : false,
"ismaster" : false,
"hidden" : false,
"secondary" : false,
"pingTimeMillis" : -1511828489
}
]
}
},
"createdByType" : {
"master" : 6,
"set" : 1
},
"totalAvailable" : 5,
"totalCreated" : 7,
"numDBClientConnection" : 8,
"numAScopedConnection" : 2,
"ok" : 1
}
One potential solution would be to use a negative sentinel value, since time must always be positive.
- duplicates
-
-
- Closed
-