Reputation: 13626
I've been trying to setup a simple replication system. 1 main mongo, 1 backup, and 1 arbiter.
Unfortunately, firing it up lead to main being elected SECONDARY, and the backup being elected PRIMARY (nice work arbiter).
Main had a priority of 100, and backup a priority of 0, along with a slave delay.
I tried to tell the backup to step down via:
PRIMARY> db.runCommand({replSetReconfig: conf})
{
"assertion" : "initiation and reconfiguration of a replica set must
be sent to a node that can become primary",
"assertionCode" : 13420,
"errmsg" : "db assertion failure",
"ok" : 0
}
PRIMARY> .adminCommand({replSetStepDown:1000000, force:1})
Fri Jan 13 17:27:29 SyntaxError: syntax error (shell):1
PRIMARY> db.adminCommand({replSetStepDown:1000000, force:1})
Fri Jan 13 17:27:36 DBClientCursor::init call() failed
Fri Jan 13 17:27:36 query failed : admin.$cmd { replSetStepDown:
1000000.0, force: 1.0 } to: 127.0.0.1
Fri Jan 13 17:27:36 Error: error doing query: failed shell/
collection.js:151
Fri Jan 13 17:27:36 trying reconnect to 127.0.0.1
Fri Jan 13 17:27:36 reconnect 127.0.0.1 ok
SECONDARY>
SECONDARY>
SECONDARY> db.adminCommand({replSetStepDown:1000000, force:1})
{ "errmsg" : "not primary so can't step down", "ok" : 0 }
Which worked, but main is still seconday as well.
Any ideas? Thanks!
config file
conf = {
version: 90002,
_id : "example",
members: [
{
_id : 1,
host : "main.example.com:27017",
priority: 100
},
{
_id : 2,
host : "backup.example.com:27017",
priority: 0,
slaveDelay : 3600
},
{
_id : 3,
host : "arbiter.example.com:27017",
priority: 0,
arbiterOnly: true
}
]
};
rs.status on main()
{
"set" : "example",
"date" : ISODate("2012-01-13T23:29:09Z"),
"myState" : 2,
"members" : [
{
"_id" : 1,
"name" : "main.example.com:27017",
"health" : 1,
"state" : 2,
"stateStr" : "SECONDARY",
"optime" : {
"t" : 1326496827000,
"i" : 1
},
"optimeDate" : ISODate("2012-01-13T23:20:27Z"),
"self" : true
},
{
"_id" : 2,
"name" : "backup.example.com:27017",
"health" : 1,
"state" : 2,
"stateStr" : "SECONDARY",
"uptime" : 324,
"optime" : {
"t" : 1326492641000,
"i" : 1
},
"optimeDate" : ISODate("2012-01-13T22:10:41Z"),
"lastHeartbeat" : ISODate("2012-01-13T23:29:09Z"),
"pingMs" : 0
},
{
"_id" : 3,
"name" : "arbiter.example.com:27017",
"health" : 1,
"state" : 7,
"stateStr" : "ARBITER",
"uptime" : 324,
"optime" : {
"t" : 0,
"i" : 0
},
"optimeDate" : ISODate("1970-01-01T00:00:00Z"),
"lastHeartbeat" : ISODate("2012-01-13T23:29:09Z"),
"pingMs" : 0
}
],
"ok" : 1
}
rs.status() on backup
{
"set" : "example",
"date" : ISODate("2012-01-13T23:31:06Z"),
"myState" : 2,
"members" : [
{
"_id" : 0,
"name" : "BACKUPVMW02:27017",
"health" : 1,
"state" : 2,
"stateStr" : "SECONDARY",
"optime" : {
"t" : 1326492641000,
"i" : 1
},
"optimeDate" : ISODate("2012-01-13T22:10:41Z"),
"self" : true
}
],
"ok" : 1
Upvotes: 3
Views: 2412
Reputation: 3320
you might be using an old version of mongodb--the field that they used for voting was called "votes" then, I think, not priority. When they switched over to the "priority" field, you should be calling rs.reconfig(configfile), I think.
Upvotes: 1
Reputation: 13626
It turns out have a 3 node setup, with 1 primary, 1 slave, and 1 arbiter doesn't work well. Removed slave delay, and priority was respected (after reinstalling all nodes from scratch and removing data directories)
Upvotes: 1