[SERVER-3204] mongodb runs on two physical nodes as replica sets, there don't get a "PRIMARY" if a host down Created: 06/Jun/11  Updated: 30/Mar/12  Resolved: 24/Jun/11

Status: Closed
Project: Core Server
Component/s: Replication
Affects Version/s: 1.8.1
Fix Version/s: None

Type: Bug Priority: Major - P3
Reporter: Linux Administrators Trust HRS Assignee: Kristina Chodorow (Inactive)
Resolution: Done Votes: 0
Labels: None
Remaining Estimate: Not Specified
Time Spent: Not Specified
Original Estimate: Not Specified
Environment:

x3850-X5, 32GB RAM, 2x Intel Xeon MP E7520; OS: Ubuntu 10.04.2 LTS, Python 2.6.5


Attachments: Text File mongodb27017.log     File rs.status     File rs.status.failed    
Operating System: Linux
Participants:

 Description   

I have two IBM server. I wont exclude single point of failure. For this I started on each server 3 mongod, two full services and one arbiter. After initializing I have 1 "primary" node, 3 "secondary" and 2 "arbiter". When I shut down a (physical) server, I have still 3 services - 1 "primary", 1 "secondary" and 1 "arbiter", but after shortly time (a few seconds) I have only 2 "secondary" and 1 "arbiter", but not "primary". What I make wrong?



 Comments   
Comment by Kristina Chodorow (Inactive) [ 30/Jun/11 ]

You only ever need 1 arbiter, but that looks fine.

Comment by Linux Administrators Trust HRS [ 28/Jun/11 ]

I have solved it that I have added a third (physical) server with an arbiter and remove from the first both servers one member. So I having now 2 servers with one active member and 1 arbiter + 1 arbiter on the additional server:
replset:PRIMARY> db.isMaster()
{
"setName" : "replset",
"ismaster" : true,
"secondary" : false,
"hosts" : [
"xx.xx.xx.73:27017",
"xx.xx.xx.74:27017"
],
"arbiters" : [
"xx.xx.xx.80:27018",
"xx.xx.xx.74:27018",
"xx.xx.xx.73:27018"
],
"maxBsonObjectSize" : 16777216,
"ok" : 1
}

Thank you.

Comment by Kristina Chodorow (Inactive) [ 24/Jun/11 ]

You need a majority of servers to be up: 3/6 is not a majority. You'd need at least 4 services out of 6 to have a majority up.

Comment by Linux Administrators Trust HRS [ 06/Jun/11 ]

rs.status() after shut down from second node

Generated at Thu Feb 08 03:02:23 UTC 2024 using Jira 9.7.1#970001-sha1:2222b88b221c4928ef0de3161136cc90c8356a66.